Update ansible-lint.yml and pylint for pub/telemetry by Kratika-P · Pull Request #4296 · dell/omnia

Kratika-P · 2026-04-21T05:13:00Z

Description of the Solution

update telemetry branch in ansible-lint workflow

Signed-off-by: Kratika Patidar <Kratika.Patidar@dell.com>

* Update openchami git version (#4251) Co-authored-by: mithileshreddy04 <mithilesh.reddy@dell.com> Co-authored-by: priti-parate <140157516+priti-parate@users.noreply.github.com> * powerscale teleemtry support with direct authentication mode * use existing vmagent * update messages in vars * merge Pub/q2 dev to pub/telemetry (#4254) * removing input template * Fix for pulp remote RemoteArtifacts is 0 after repo migration Signed-off-by: pullan1 <sudha.pullalaravu@dell.com> --------- Signed-off-by: pullan1 <sudha.pullalaravu@dell.com> Co-authored-by: pullan1 <sudha.pullalaravu@dell.com> Co-authored-by: snarthan <narthan.s@dell.com> * Powerscale teleemtry support using helm * deploy powerscale telemetry using cloud-init * offline deployment of powerscale telemetry * fix for cert-manager failure * fix for cert manager failure * powerscale telemetry deployment with telemetry namespace * sync q2_dev changes (#4263) * removing input template * Fix for pulp remote RemoteArtifacts is 0 after repo migration Signed-off-by: pullan1 <sudha.pullalaravu@dell.com> * Feature/ome discovery pxe mapping enhancements (#4245) * feat(discovery): OME static group extraction, PXE mapping IP/SU/parent tag enhancements ome_server_inventory.py: - Fix static group extraction: find 'Static Groups' container by name and select only direct children via ParentId; avoids picking system/nested groups - Emit module.warn() for static groups that exist but have no devices assigned - Fix idrac_hostname: read InstrumentationName/DnsName from DeviceManagement ManagementType==2 entry instead of DeviceName which returns the IP address generate_pxe_mapping.py: - ADMIN_IP: derive from first 2 octets of admin_network.subnet + last 2 of BMC IP - IB_IP: derive from first 2 octets of ib_network.subnet + last 2 of BMC IP - Skip IB_IP/IB_MAC when server has no IB NIC (ib_nic_mac is empty) - Add extract_su_from_hostname() with regex (SU[A-Z]?\d+)(?=R\d+) to parse Scalable Unit from BMC hostname; rejects service-tag-only hostnames (idrac-JCGT033) and falls back to grp0 when no SU pattern is found - Set GROUP_NAME to extracted SU identifier (fallback: grp0) - Post-process rows to assign PARENT_SERVICE_TAG from the service_kube_control_plane_x86_64 node within the same SU group - Remove BMC_HOSTNAME from CSV headers and output rows - Lint: remove dead try/except in calculate_admin_ip/calculate_ib_ip, reuse ib_mac variable, suppress broad-except pylint warning generate_pxe_mapping.yml: - Load network_spec.yml via include_vars - Set admin_subnet and ib_subnet using selectattr on Networks list - Pass both subnets as parameters to the generate_pxe_mapping module defaults/main.yml: - Add admin_subnet and ib_subnet default variables (empty string) provision_validation.py: - Comment out validate_admin_ips_against_network_spec function and its call site; ADMIN_IPs are now derived from subnet octets + BMC IP and will not necessarily fall within primary_oim_admin_ip/netmask_bits range * refactor: rename discovery directory to provision, update network_spec.yml - Renamed discovery/ to provision/ (git detected as rename, no content loss) - Updated input/network_spec.yml with latest network configuration changes * Update discovery.yml * refactor: unify OME credentials into get_config_credentials flow - Added ome_ip, ome_username, ome_password to omnia_credential.j2 template - Added 'discovery' service entry to omnia_credentials in update_config/vars/main.yml - Added 'discovery' to the hardcoded service key trigger list in fetch_credentials.yml - Replaced custom vault logic in get_ome_credentials.yml with unified decrypt_include_encrypt.yml call against omnia_config_credentials.yml - Updated ome_discovery/vars/main.yml to reference omnia_config_credentials_file and omnia_config_credentials_vault_key instead of the separate .vault/ paths - Deleted .vault/ome_credentials.yml and .vault/.vault_password (no longer needed) * chore: update copyright year from 2025 to 2026 in modified files Updated copyright header in all ome_discovery files modified during this feature branch: - library/generate_pxe_mapping.py - library/ome_server_inventory.py - tasks/generate_pxe_mapping.yml - tasks/get_ome_credentials.yml - defaults/main.yml - vars/main.yml * fix: restore discovery_validations role missed during discovery-to-provision rename discovery/roles/discovery_validations/ was accidentally dropped when renaming the discovery/ directory to provision/. Add it back under provision/roles/discovery_validations/ to resolve the PR merge conflict. * chore: update copyright year to 2026 in provision/roles/discovery_validations files * fix: remove duplicate discovery_validations role (provision_validations already exists) provision/roles/provision_validations/ is the correct renamed equivalent of discovery/roles/discovery_validations/. The discovery_validations copy added to provision/ was redundant. * feat: apply upstream telemetry upgrade changes from dell/omnia pub/q2_dev - Replace kubectl command with kubernetes.core.k8s module for iDRAC StatefulSet - Preserve existing replica count during iDRAC StatefulSet upgrade - Add LDMS store daemon check, restart, and readiness wait tasks * fix: quote build_stream_job_id_absent message in provision_validations vars * feat: add discovery/roles/discovery_validations and telemetry files - Add discovery/roles/discovery_validations/vars/main.yml with task definitions for validation flow - Add discovery/roles/telemetry/tasks/apply_telemetry_on_upgrade.yml with upstream telemetry upgrade logic (replica preservation + LDMS store) * fix: wrap long line in fetch_credentials.yml to satisfy yaml[line-length] lint * refactor: move ome_ip from credentials to discovery_config.yml - Create input/discovery_config.yml for non-credential discovery settings (ome_ip, future Magellan config) - Remove ome_ip from omnia_credential.j2 and credential update vars - Load ome_ip via include_vars from discovery_config.yml in get_ome_credentials.yml - Add discovery_config.yml to provision_validations discovery_inputs - Remove redundant ib_subnet/admin_subnet defaults from ome_discovery * fix: add newline at end of ome_discovery/defaults/main.yml * fix: override role_path to absolute path for decrypt_include_encrypt.yml role_path resolves to ome_discovery role path, causing encrypt_files_vars.yml to be looked up incorrectly. Override to playbook_dir dirname (/opt/omnia/omnia). * fix: inline credential loading to avoid role_path resolution issue role_path cannot be overridden in include_tasks vars. Replace the call to decrypt_include_encrypt.yml with direct include_vars using stat checks for encrypted vs unencrypted credential file handling. * fix: skip load-failure rule in ansible-lint to avoid CI false positives ansible-lint fails to resolve role_path relative paths during static analysis in GitHub Actions, causing false load-failure errors for files that exist and work at runtime. * Update ansible.cfg * Update ansible.cfg * refactor: rename discovery references to provision and add discovery_config variable - Rename discover_mapping_nodes.yml to provision_mapping_nodes.yml - Replace "discovery" terminology with "provision" across playbooks, vars, READMEs, and task names in provision roles - Add subnet as required field with IP pattern validation in network_spec schema - Define discovery_config variable in ome_discovery vars and use it in get_ome_credentials.yml (consistent with provision_config pattern) - Rename discovery_inputs to provision_inputs in validation vars - Rename discovery_mech_mapping to provision_mech_mapping - Update user-facing messages to reference provision.yml Generated with [Devin](https://cli.devin.ai/docs) Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> * fix: credential rules, vault handling, GROUP_NAME validation, and discovery playbook improvements - Add ome_username and ome_password validation rules to credential_rules.json - Add 'discovery' tag to prepare_oim omnia_run_tags so OME credentials are prompted - Fix vault-encrypted credential loading in get_ome_credentials.yml (use decrypt-include-reencrypt pattern instead of unsupported vault_password_file) - Add include_input_dir.yml import to discovery.yml so input_project_dir is set - Accept SU1-SU100 (case-insensitive) in addition to grp0-grp100 for GROUP_NAME - Fix Magellan message to use list format (avoids \n in debug output) - Remove escaped quotes from discovery usage examples Generated with [Devin](https://cli.devin.ai/docs) Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> * fix: extend SU group name support to build_image validation and schemas - Add build_aarch_image tag to input_file_inventory so build_image_aarch64.yml runs provision_config validation (was missing, causing no validation to run for aarch64 builds) - Update GROUP_NAME patterns in functional_groups_config.json and omnia_config.json schemas to accept SU1-SU100 format alongside grp0-grp100 - Update INVALID_GROUP_NAME_MSG to reflect both accepted formats Generated with [Devin](https://cli.devin.ai/docs) Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> --------- Signed-off-by: Sujit Jadhav <sujit_jadhav@dell.com> Co-authored-by: Super User <root@testbed.omnia.test> Co-authored-by: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> * Cleanup discovery roles: move library modules, remove unused roles (#4261) * Cleanup discovery roles: move library modules, remove unused roles - Move ome_server_inventory.py and generate_pxe_mapping.py from discovery/roles/ome_discovery/library/ to common/library/modules/ so they are shared via the common module search path already configured in discovery/ansible.cfg - Remove unused discovery/roles/telemetry/ directory - Remove unused discovery/roles/discovery_validations/ directory - Load discovery_config.yml at playbook level in discovery.yml (consistent with how build_stream_config.yml is loaded in provision.yml) - Fix discovery_complete_msg formatting for readable Ansible output Generated with [Devin](https://cli.devin.ai/docs) Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> * Remove unused discovery_validations role Generated with [Devin](https://cli.devin.ai/docs) Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> --------- Co-authored-by: Super User <root@testbed.omnia.test> Co-authored-by: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> * fix for set_pxe_boot.yml when custom inventory given (#4260) * Update generate_bmc_inventory.yml Signed-off-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> * Update pre_checks.yml Signed-off-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> * lint issue Signed-off-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> --------- Signed-off-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> --------- Signed-off-by: pullan1 <sudha.pullalaravu@dell.com> Signed-off-by: Sujit Jadhav <sujit_jadhav@dell.com> Signed-off-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> Co-authored-by: pullan1 <sudha.pullalaravu@dell.com> Co-authored-by: snarthan <narthan.s@dell.com> Co-authored-by: Sujit Jadhav <sujit_jadhav@dell.com> Co-authored-by: Super User <root@testbed.omnia.test> Co-authored-by: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> Co-authored-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> * resolving merge conflict * revert openchami commit id * resolving review comments * addressing review comments * fix for vmagent scraping powerscale metrics * cleanup script correction for powerscale telemetry cleanup * victoria operator and victoria log input validation * vitoria log input and input validation * remving L2 vslidation for victoria log which is not required * input validation and review comment addressing * change idrac_telemetry_collection_type to telemetry_collection_type * Remove invisible Unicode LRM (U+200E) characters from victoria-operator template filenames * VictoriaLogs container image references and default variable * port check * resolve merge conflict * correction for schema * Update telemetry_config.json * Update validate_input.py * merge conflict telemetry_prereq.yml * change victoria_configurations to victoria_metrics_configurations * remove deployment mode input variable * update for upgrade scenarios * update comments * update comment * resolving issues due to merge conflict * vitoria log changes * victoria log cluster component and VLAgent deployment * updating pod name * removing the changes of adding cert * victoria log changes * remivng victoria log pod calidation playbook * cleanup changes for victoria log * Update ansible-lint.yml and pylint for pub/telemetry (#4296) * Update ansible-lint.yml Signed-off-by: Kratika Patidar <Kratika.Patidar@dell.com> * Update pylint.yml Signed-off-by: Kratika Patidar <Kratika.Patidar@dell.com> * fixing ansible-lint * lint * line-lenght --------- Signed-off-by: Kratika Patidar <Kratika.Patidar@dell.com> --------- Signed-off-by: pullan1 <sudha.pullalaravu@dell.com> Signed-off-by: Sujit Jadhav <sujit_jadhav@dell.com> Signed-off-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> Signed-off-by: priti-parate <140157516+priti-parate@users.noreply.github.com> Signed-off-by: Kratika Patidar <Kratika.Patidar@dell.com> Co-authored-by: mithileshreddy04 <mithilesh.reddy@dell.com> Co-authored-by: priti-parate <140157516+priti-parate@users.noreply.github.com> Co-authored-by: pullan1 <sudha.pullalaravu@dell.com> Co-authored-by: snarthan <narthan.s@dell.com> Co-authored-by: Sujit Jadhav <sujit_jadhav@dell.com> Co-authored-by: Super User <root@testbed.omnia.test> Co-authored-by: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> Co-authored-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> Co-authored-by: Kratika_Patidar <Kratika.Patidar@dell.com>

* Merge pull request #4294 from mithileshreddy04/pub/q2_dev OpenCHAMI upgrade changes in prepare_oim and oim_cleanup * Feature branch sync - pub/telemetry to pub/q2_dev (#4293) * Update openchami git version (#4251) Co-authored-by: mithileshreddy04 <mithilesh.reddy@dell.com> Co-authored-by: priti-parate <140157516+priti-parate@users.noreply.github.com> * powerscale teleemtry support with direct authentication mode * use existing vmagent * update messages in vars * merge Pub/q2 dev to pub/telemetry (#4254) * removing input template * Fix for pulp remote RemoteArtifacts is 0 after repo migration Signed-off-by: pullan1 <sudha.pullalaravu@dell.com> --------- Signed-off-by: pullan1 <sudha.pullalaravu@dell.com> Co-authored-by: pullan1 <sudha.pullalaravu@dell.com> Co-authored-by: snarthan <narthan.s@dell.com> * Powerscale teleemtry support using helm * deploy powerscale telemetry using cloud-init * offline deployment of powerscale telemetry * fix for cert-manager failure * fix for cert manager failure * powerscale telemetry deployment with telemetry namespace * sync q2_dev changes (#4263) * removing input template * Fix for pulp remote RemoteArtifacts is 0 after repo migration Signed-off-by: pullan1 <sudha.pullalaravu@dell.com> * Feature/ome discovery pxe mapping enhancements (#4245) * feat(discovery): OME static group extraction, PXE mapping IP/SU/parent tag enhancements ome_server_inventory.py: - Fix static group extraction: find 'Static Groups' container by name and select only direct children via ParentId; avoids picking system/nested groups - Emit module.warn() for static groups that exist but have no devices assigned - Fix idrac_hostname: read InstrumentationName/DnsName from DeviceManagement ManagementType==2 entry instead of DeviceName which returns the IP address generate_pxe_mapping.py: - ADMIN_IP: derive from first 2 octets of admin_network.subnet + last 2 of BMC IP - IB_IP: derive from first 2 octets of ib_network.subnet + last 2 of BMC IP - Skip IB_IP/IB_MAC when server has no IB NIC (ib_nic_mac is empty) - Add extract_su_from_hostname() with regex (SU[A-Z]?\d+)(?=R\d+) to parse Scalable Unit from BMC hostname; rejects service-tag-only hostnames (idrac-JCGT033) and falls back to grp0 when no SU pattern is found - Set GROUP_NAME to extracted SU identifier (fallback: grp0) - Post-process rows to assign PARENT_SERVICE_TAG from the service_kube_control_plane_x86_64 node within the same SU group - Remove BMC_HOSTNAME from CSV headers and output rows - Lint: remove dead try/except in calculate_admin_ip/calculate_ib_ip, reuse ib_mac variable, suppress broad-except pylint warning generate_pxe_mapping.yml: - Load network_spec.yml via include_vars - Set admin_subnet and ib_subnet using selectattr on Networks list - Pass both subnets as parameters to the generate_pxe_mapping module defaults/main.yml: - Add admin_subnet and ib_subnet default variables (empty string) provision_validation.py: - Comment out validate_admin_ips_against_network_spec function and its call site; ADMIN_IPs are now derived from subnet octets + BMC IP and will not necessarily fall within primary_oim_admin_ip/netmask_bits range * refactor: rename discovery directory to provision, update network_spec.yml - Renamed discovery/ to provision/ (git detected as rename, no content loss) - Updated input/network_spec.yml with latest network configuration changes * Update discovery.yml * refactor: unify OME credentials into get_config_credentials flow - Added ome_ip, ome_username, ome_password to omnia_credential.j2 template - Added 'discovery' service entry to omnia_credentials in update_config/vars/main.yml - Added 'discovery' to the hardcoded service key trigger list in fetch_credentials.yml - Replaced custom vault logic in get_ome_credentials.yml with unified decrypt_include_encrypt.yml call against omnia_config_credentials.yml - Updated ome_discovery/vars/main.yml to reference omnia_config_credentials_file and omnia_config_credentials_vault_key instead of the separate .vault/ paths - Deleted .vault/ome_credentials.yml and .vault/.vault_password (no longer needed) * chore: update copyright year from 2025 to 2026 in modified files Updated copyright header in all ome_discovery files modified during this feature branch: - library/generate_pxe_mapping.py - library/ome_server_inventory.py - tasks/generate_pxe_mapping.yml - tasks/get_ome_credentials.yml - defaults/main.yml - vars/main.yml * fix: restore discovery_validations role missed during discovery-to-provision rename discovery/roles/discovery_validations/ was accidentally dropped when renaming the discovery/ directory to provision/. Add it back under provision/roles/discovery_validations/ to resolve the PR merge conflict. * chore: update copyright year to 2026 in provision/roles/discovery_validations files * fix: remove duplicate discovery_validations role (provision_validations already exists) provision/roles/provision_validations/ is the correct renamed equivalent of discovery/roles/discovery_validations/. The discovery_validations copy added to provision/ was redundant. * feat: apply upstream telemetry upgrade changes from dell/omnia pub/q2_dev - Replace kubectl command with kubernetes.core.k8s module for iDRAC StatefulSet - Preserve existing replica count during iDRAC StatefulSet upgrade - Add LDMS store daemon check, restart, and readiness wait tasks * fix: quote build_stream_job_id_absent message in provision_validations vars * feat: add discovery/roles/discovery_validations and telemetry files - Add discovery/roles/discovery_validations/vars/main.yml with task definitions for validation flow - Add discovery/roles/telemetry/tasks/apply_telemetry_on_upgrade.yml with upstream telemetry upgrade logic (replica preservation + LDMS store) * fix: wrap long line in fetch_credentials.yml to satisfy yaml[line-length] lint * refactor: move ome_ip from credentials to discovery_config.yml - Create input/discovery_config.yml for non-credential discovery settings (ome_ip, future Magellan config) - Remove ome_ip from omnia_credential.j2 and credential update vars - Load ome_ip via include_vars from discovery_config.yml in get_ome_credentials.yml - Add discovery_config.yml to provision_validations discovery_inputs - Remove redundant ib_subnet/admin_subnet defaults from ome_discovery * fix: add newline at end of ome_discovery/defaults/main.yml * fix: override role_path to absolute path for decrypt_include_encrypt.yml role_path resolves to ome_discovery role path, causing encrypt_files_vars.yml to be looked up incorrectly. Override to playbook_dir dirname (/opt/omnia/omnia). * fix: inline credential loading to avoid role_path resolution issue role_path cannot be overridden in include_tasks vars. Replace the call to decrypt_include_encrypt.yml with direct include_vars using stat checks for encrypted vs unencrypted credential file handling. * fix: skip load-failure rule in ansible-lint to avoid CI false positives ansible-lint fails to resolve role_path relative paths during static analysis in GitHub Actions, causing false load-failure errors for files that exist and work at runtime. * Update ansible.cfg * Update ansible.cfg * refactor: rename discovery references to provision and add discovery_config variable - Rename discover_mapping_nodes.yml to provision_mapping_nodes.yml - Replace "discovery" terminology with "provision" across playbooks, vars, READMEs, and task names in provision roles - Add subnet as required field with IP pattern validation in network_spec schema - Define discovery_config variable in ome_discovery vars and use it in get_ome_credentials.yml (consistent with provision_config pattern) - Rename discovery_inputs to provision_inputs in validation vars - Rename discovery_mech_mapping to provision_mech_mapping - Update user-facing messages to reference provision.yml Generated with [Devin](https://cli.devin.ai/docs) Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> * fix: credential rules, vault handling, GROUP_NAME validation, and discovery playbook improvements - Add ome_username and ome_password validation rules to credential_rules.json - Add 'discovery' tag to prepare_oim omnia_run_tags so OME credentials are prompted - Fix vault-encrypted credential loading in get_ome_credentials.yml (use decrypt-include-reencrypt pattern instead of unsupported vault_password_file) - Add include_input_dir.yml import to discovery.yml so input_project_dir is set - Accept SU1-SU100 (case-insensitive) in addition to grp0-grp100 for GROUP_NAME - Fix Magellan message to use list format (avoids \n in debug output) - Remove escaped quotes from discovery usage examples Generated with [Devin](https://cli.devin.ai/docs) Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> * fix: extend SU group name support to build_image validation and schemas - Add build_aarch_image tag to input_file_inventory so build_image_aarch64.yml runs provision_config validation (was missing, causing no validation to run for aarch64 builds) - Update GROUP_NAME patterns in functional_groups_config.json and omnia_config.json schemas to accept SU1-SU100 format alongside grp0-grp100 - Update INVALID_GROUP_NAME_MSG to reflect both accepted formats Generated with [Devin](https://cli.devin.ai/docs) Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> --------- Signed-off-by: Sujit Jadhav <sujit_jadhav@dell.com> Co-authored-by: Super User <root@testbed.omnia.test> Co-authored-by: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> * Cleanup discovery roles: move library modules, remove unused roles (#4261) * Cleanup discovery roles: move library modules, remove unused roles - Move ome_server_inventory.py and generate_pxe_mapping.py from discovery/roles/ome_discovery/library/ to common/library/modules/ so they are shared via the common module search path already configured in discovery/ansible.cfg - Remove unused discovery/roles/telemetry/ directory - Remove unused discovery/roles/discovery_validations/ directory - Load discovery_config.yml at playbook level in discovery.yml (consistent with how build_stream_config.yml is loaded in provision.yml) - Fix discovery_complete_msg formatting for readable Ansible output Generated with [Devin](https://cli.devin.ai/docs) Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> * Remove unused discovery_validations role Generated with [Devin](https://cli.devin.ai/docs) Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> --------- Co-authored-by: Super User <root@testbed.omnia.test> Co-authored-by: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> * fix for set_pxe_boot.yml when custom inventory given (#4260) * Update generate_bmc_inventory.yml Signed-off-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> * Update pre_checks.yml Signed-off-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> * lint issue Signed-off-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> --------- Signed-off-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> --------- Signed-off-by: pullan1 <sudha.pullalaravu@dell.com> Signed-off-by: Sujit Jadhav <sujit_jadhav@dell.com> Signed-off-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> Co-authored-by: pullan1 <sudha.pullalaravu@dell.com> Co-authored-by: snarthan <narthan.s@dell.com> Co-authored-by: Sujit Jadhav <sujit_jadhav@dell.com> Co-authored-by: Super User <root@testbed.omnia.test> Co-authored-by: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> Co-authored-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> * resolving merge conflict * revert openchami commit id * resolving review comments * addressing review comments * fix for vmagent scraping powerscale metrics * cleanup script correction for powerscale telemetry cleanup * victoria operator and victoria log input validation * vitoria log input and input validation * remving L2 vslidation for victoria log which is not required * input validation and review comment addressing * change idrac_telemetry_collection_type to telemetry_collection_type * Remove invisible Unicode LRM (U+200E) characters from victoria-operator template filenames * VictoriaLogs container image references and default variable * port check * resolve merge conflict * correction for schema * Update telemetry_config.json * Update validate_input.py * merge conflict telemetry_prereq.yml * change victoria_configurations to victoria_metrics_configurations * remove deployment mode input variable * update for upgrade scenarios * update comments * update comment * resolving issues due to merge conflict * vitoria log changes * victoria log cluster component and VLAgent deployment * updating pod name * removing the changes of adding cert * victoria log changes * remivng victoria log pod calidation playbook * cleanup changes for victoria log * Update ansible-lint.yml and pylint for pub/telemetry (#4296) * Update ansible-lint.yml Signed-off-by: Kratika Patidar <Kratika.Patidar@dell.com> * Update pylint.yml Signed-off-by: Kratika Patidar <Kratika.Patidar@dell.com> * fixing ansible-lint * lint * line-lenght --------- Signed-off-by: Kratika Patidar <Kratika.Patidar@dell.com> --------- Signed-off-by: pullan1 <sudha.pullalaravu@dell.com> Signed-off-by: Sujit Jadhav <sujit_jadhav@dell.com> Signed-off-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> Signed-off-by: priti-parate <140157516+priti-parate@users.noreply.github.com> Signed-off-by: Kratika Patidar <Kratika.Patidar@dell.com> Co-authored-by: mithileshreddy04 <mithilesh.reddy@dell.com> Co-authored-by: priti-parate <140157516+priti-parate@users.noreply.github.com> Co-authored-by: pullan1 <sudha.pullalaravu@dell.com> Co-authored-by: snarthan <narthan.s@dell.com> Co-authored-by: Sujit Jadhav <sujit_jadhav@dell.com> Co-authored-by: Super User <root@testbed.omnia.test> Co-authored-by: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> Co-authored-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> Co-authored-by: Kratika_Patidar <Kratika.Patidar@dell.com> * IB nic ip assignment * update MinIO and registry images to fixed tagged versions, omnia core container tag and version to 2.2 and v2.2.0.0 (#4309) * Minimal OS-only functional group enablement for x86_64 and aarch64 * Update image_package_collector.py * Update provision_validation.py * Minimal OS functional group updates in provision * Minimal OS functional group upgrade * Fix os_* package cross-contamination and remove stale discovery templates * OpenCHAMI upgrade changes * Update openchami container tags * Update main.yml * Update main.yml * Update main.yml * Update omnia version and core tag --------- Signed-off-by: pullan1 <sudha.pullalaravu@dell.com> Signed-off-by: Sujit Jadhav <sujit_jadhav@dell.com> Signed-off-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> Signed-off-by: priti-parate <140157516+priti-parate@users.noreply.github.com> Signed-off-by: Kratika Patidar <Kratika.Patidar@dell.com> Co-authored-by: Mithilesh Reddy <mithilesh.reddy@dell.com> Co-authored-by: priti-parate <140157516+priti-parate@users.noreply.github.com> Co-authored-by: pullan1 <sudha.pullalaravu@dell.com> Co-authored-by: snarthan <narthan.s@dell.com> Co-authored-by: Sujit Jadhav <sujit_jadhav@dell.com> Co-authored-by: Super User <root@testbed.omnia.test> Co-authored-by: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> Co-authored-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> Co-authored-by: Kratika_Patidar <Kratika.Patidar@dell.com> Co-authored-by: Nagachandan-P <Nagachandan.p@dell.com>

* Minimal OS-only functional group enablement (#4267) * Minimal OS-only functional group enablement for x86_64 and aarch64 * Update image_package_collector.py * Update provision_validation.py * Minimal OS functional group updates in provision * Minimal OS functional group upgrade * Fix os_* package cross-contamination and remove stale discovery templates * csi defect fix * nvidia dcgm install * Add 'provision' tag to omnia_run_tags Signed-off-by: balajikumaran.cs <balajikumaran.c.s@gmail.com> * Add 'provision' tag to omnia_run_tags (#4276) Signed-off-by: balajikumaran.cs <balajikumaran.c.s@gmail.com> * Fix incorrect file path in Podman login failure message (OMN01D-2166) The podman_login_fail_msg referenced a hardcoded path input/omnia_config_credentials.yml which does not exist. Updated it to use the dynamic input_project_dir variable so the error message now correctly points to the actual credentials file at <input_project_dir>/omnia_config_credentials.yml. Fixed in both prepare_oim and gitlab roles. * Merge pull request #4294 from mithileshreddy04/pub/q2_dev OpenCHAMI upgrade changes in prepare_oim and oim_cleanup * Feature branch sync - pub/telemetry to pub/q2_dev (#4293) * Update openchami git version (#4251) Co-authored-by: mithileshreddy04 <mithilesh.reddy@dell.com> Co-authored-by: priti-parate <140157516+priti-parate@users.noreply.github.com> * powerscale teleemtry support with direct authentication mode * use existing vmagent * update messages in vars * merge Pub/q2 dev to pub/telemetry (#4254) * removing input template * Fix for pulp remote RemoteArtifacts is 0 after repo migration Signed-off-by: pullan1 <sudha.pullalaravu@dell.com> --------- Signed-off-by: pullan1 <sudha.pullalaravu@dell.com> Co-authored-by: pullan1 <sudha.pullalaravu@dell.com> Co-authored-by: snarthan <narthan.s@dell.com> * Powerscale teleemtry support using helm * deploy powerscale telemetry using cloud-init * offline deployment of powerscale telemetry * fix for cert-manager failure * fix for cert manager failure * powerscale telemetry deployment with telemetry namespace * sync q2_dev changes (#4263) * removing input template * Fix for pulp remote RemoteArtifacts is 0 after repo migration Signed-off-by: pullan1 <sudha.pullalaravu@dell.com> * Feature/ome discovery pxe mapping enhancements (#4245) * feat(discovery): OME static group extraction, PXE mapping IP/SU/parent tag enhancements ome_server_inventory.py: - Fix static group extraction: find 'Static Groups' container by name and select only direct children via ParentId; avoids picking system/nested groups - Emit module.warn() for static groups that exist but have no devices assigned - Fix idrac_hostname: read InstrumentationName/DnsName from DeviceManagement ManagementType==2 entry instead of DeviceName which returns the IP address generate_pxe_mapping.py: - ADMIN_IP: derive from first 2 octets of admin_network.subnet + last 2 of BMC IP - IB_IP: derive from first 2 octets of ib_network.subnet + last 2 of BMC IP - Skip IB_IP/IB_MAC when server has no IB NIC (ib_nic_mac is empty) - Add extract_su_from_hostname() with regex (SU[A-Z]?\d+)(?=R\d+) to parse Scalable Unit from BMC hostname; rejects service-tag-only hostnames (idrac-JCGT033) and falls back to grp0 when no SU pattern is found - Set GROUP_NAME to extracted SU identifier (fallback: grp0) - Post-process rows to assign PARENT_SERVICE_TAG from the service_kube_control_plane_x86_64 node within the same SU group - Remove BMC_HOSTNAME from CSV headers and output rows - Lint: remove dead try/except in calculate_admin_ip/calculate_ib_ip, reuse ib_mac variable, suppress broad-except pylint warning generate_pxe_mapping.yml: - Load network_spec.yml via include_vars - Set admin_subnet and ib_subnet using selectattr on Networks list - Pass both subnets as parameters to the generate_pxe_mapping module defaults/main.yml: - Add admin_subnet and ib_subnet default variables (empty string) provision_validation.py: - Comment out validate_admin_ips_against_network_spec function and its call site; ADMIN_IPs are now derived from subnet octets + BMC IP and will not necessarily fall within primary_oim_admin_ip/netmask_bits range * refactor: rename discovery directory to provision, update network_spec.yml - Renamed discovery/ to provision/ (git detected as rename, no content loss) - Updated input/network_spec.yml with latest network configuration changes * Update discovery.yml * refactor: unify OME credentials into get_config_credentials flow - Added ome_ip, ome_username, ome_password to omnia_credential.j2 template - Added 'discovery' service entry to omnia_credentials in update_config/vars/main.yml - Added 'discovery' to the hardcoded service key trigger list in fetch_credentials.yml - Replaced custom vault logic in get_ome_credentials.yml with unified decrypt_include_encrypt.yml call against omnia_config_credentials.yml - Updated ome_discovery/vars/main.yml to reference omnia_config_credentials_file and omnia_config_credentials_vault_key instead of the separate .vault/ paths - Deleted .vault/ome_credentials.yml and .vault/.vault_password (no longer needed) * chore: update copyright year from 2025 to 2026 in modified files Updated copyright header in all ome_discovery files modified during this feature branch: - library/generate_pxe_mapping.py - library/ome_server_inventory.py - tasks/generate_pxe_mapping.yml - tasks/get_ome_credentials.yml - defaults/main.yml - vars/main.yml * fix: restore discovery_validations role missed during discovery-to-provision rename discovery/roles/discovery_validations/ was accidentally dropped when renaming the discovery/ directory to provision/. Add it back under provision/roles/discovery_validations/ to resolve the PR merge conflict. * chore: update copyright year to 2026 in provision/roles/discovery_validations files * fix: remove duplicate discovery_validations role (provision_validations already exists) provision/roles/provision_validations/ is the correct renamed equivalent of discovery/roles/discovery_validations/. The discovery_validations copy added to provision/ was redundant. * feat: apply upstream telemetry upgrade changes from dell/omnia pub/q2_dev - Replace kubectl command with kubernetes.core.k8s module for iDRAC StatefulSet - Preserve existing replica count during iDRAC StatefulSet upgrade - Add LDMS store daemon check, restart, and readiness wait tasks * fix: quote build_stream_job_id_absent message in provision_validations vars * feat: add discovery/roles/discovery_validations and telemetry files - Add discovery/roles/discovery_validations/vars/main.yml with task definitions for validation flow - Add discovery/roles/telemetry/tasks/apply_telemetry_on_upgrade.yml with upstream telemetry upgrade logic (replica preservation + LDMS store) * fix: wrap long line in fetch_credentials.yml to satisfy yaml[line-length] lint * refactor: move ome_ip from credentials to discovery_config.yml - Create input/discovery_config.yml for non-credential discovery settings (ome_ip, future Magellan config) - Remove ome_ip from omnia_credential.j2 and credential update vars - Load ome_ip via include_vars from discovery_config.yml in get_ome_credentials.yml - Add discovery_config.yml to provision_validations discovery_inputs - Remove redundant ib_subnet/admin_subnet defaults from ome_discovery * fix: add newline at end of ome_discovery/defaults/main.yml * fix: override role_path to absolute path for decrypt_include_encrypt.yml role_path resolves to ome_discovery role path, causing encrypt_files_vars.yml to be looked up incorrectly. Override to playbook_dir dirname (/opt/omnia/omnia). * fix: inline credential loading to avoid role_path resolution issue role_path cannot be overridden in include_tasks vars. Replace the call to decrypt_include_encrypt.yml with direct include_vars using stat checks for encrypted vs unencrypted credential file handling. * fix: skip load-failure rule in ansible-lint to avoid CI false positives ansible-lint fails to resolve role_path relative paths during static analysis in GitHub Actions, causing false load-failure errors for files that exist and work at runtime. * Update ansible.cfg * Update ansible.cfg * refactor: rename discovery references to provision and add discovery_config variable - Rename discover_mapping_nodes.yml to provision_mapping_nodes.yml - Replace "discovery" terminology with "provision" across playbooks, vars, READMEs, and task names in provision roles - Add subnet as required field with IP pattern validation in network_spec schema - Define discovery_config variable in ome_discovery vars and use it in get_ome_credentials.yml (consistent with provision_config pattern) - Rename discovery_inputs to provision_inputs in validation vars - Rename discovery_mech_mapping to provision_mech_mapping - Update user-facing messages to reference provision.yml Generated with [Devin](https://cli.devin.ai/docs) Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> * fix: credential rules, vault handling, GROUP_NAME validation, and discovery playbook improvements - Add ome_username and ome_password validation rules to credential_rules.json - Add 'discovery' tag to prepare_oim omnia_run_tags so OME credentials are prompted - Fix vault-encrypted credential loading in get_ome_credentials.yml (use decrypt-include-reencrypt pattern instead of unsupported vault_password_file) - Add include_input_dir.yml import to discovery.yml so input_project_dir is set - Accept SU1-SU100 (case-insensitive) in addition to grp0-grp100 for GROUP_NAME - Fix Magellan message to use list format (avoids \n in debug output) - Remove escaped quotes from discovery usage examples Generated with [Devin](https://cli.devin.ai/docs) Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> * fix: extend SU group name support to build_image validation and schemas - Add build_aarch_image tag to input_file_inventory so build_image_aarch64.yml runs provision_config validation (was missing, causing no validation to run for aarch64 builds) - Update GROUP_NAME patterns in functional_groups_config.json and omnia_config.json schemas to accept SU1-SU100 format alongside grp0-grp100 - Update INVALID_GROUP_NAME_MSG to reflect both accepted formats Generated with [Devin](https://cli.devin.ai/docs) Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> --------- Signed-off-by: Sujit Jadhav <sujit_jadhav@dell.com> Co-authored-by: Super User <root@testbed.omnia.test> Co-authored-by: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> * Cleanup discovery roles: move library modules, remove unused roles (#4261) * Cleanup discovery roles: move library modules, remove unused roles - Move ome_server_inventory.py and generate_pxe_mapping.py from discovery/roles/ome_discovery/library/ to common/library/modules/ so they are shared via the common module search path already configured in discovery/ansible.cfg - Remove unused discovery/roles/telemetry/ directory - Remove unused discovery/roles/discovery_validations/ directory - Load discovery_config.yml at playbook level in discovery.yml (consistent with how build_stream_config.yml is loaded in provision.yml) - Fix discovery_complete_msg formatting for readable Ansible output Generated with [Devin](https://cli.devin.ai/docs) Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> * Remove unused discovery_validations role Generated with [Devin](https://cli.devin.ai/docs) Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> --------- Co-authored-by: Super User <root@testbed.omnia.test> Co-authored-by: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> * fix for set_pxe_boot.yml when custom inventory given (#4260) * Update generate_bmc_inventory.yml Signed-off-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> * Update pre_checks.yml Signed-off-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> * lint issue Signed-off-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> --------- Signed-off-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> --------- Signed-off-by: pullan1 <sudha.pullalaravu@dell.com> Signed-off-by: Sujit Jadhav <sujit_jadhav@dell.com> Signed-off-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> Co-authored-by: pullan1 <sudha.pullalaravu@dell.com> Co-authored-by: snarthan <narthan.s@dell.com> Co-authored-by: Sujit Jadhav <sujit_jadhav@dell.com> Co-authored-by: Super User <root@testbed.omnia.test> Co-authored-by: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> Co-authored-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> * resolving merge conflict * revert openchami commit id * resolving review comments * addressing review comments * fix for vmagent scraping powerscale metrics * cleanup script correction for powerscale telemetry cleanup * victoria operator and victoria log input validation * vitoria log input and input validation * remving L2 vslidation for victoria log which is not required * input validation and review comment addressing * change idrac_telemetry_collection_type to telemetry_collection_type * Remove invisible Unicode LRM (U+200E) characters from victoria-operator template filenames * VictoriaLogs container image references and default variable * port check * resolve merge conflict * correction for schema * Update telemetry_config.json * Update validate_input.py * merge conflict telemetry_prereq.yml * change victoria_configurations to victoria_metrics_configurations * remove deployment mode input variable * update for upgrade scenarios * update comments * update comment * resolving issues due to merge conflict * vitoria log changes * victoria log cluster component and VLAgent deployment * updating pod name * removing the changes of adding cert * victoria log changes * remivng victoria log pod calidation playbook * cleanup changes for victoria log * Update ansible-lint.yml and pylint for pub/telemetry (#4296) * Update ansible-lint.yml Signed-off-by: Kratika Patidar <Kratika.Patidar@dell.com> * Update pylint.yml Signed-off-by: Kratika Patidar <Kratika.Patidar@dell.com> * fixing ansible-lint * lint * line-lenght --------- Signed-off-by: Kratika Patidar <Kratika.Patidar@dell.com> --------- Signed-off-by: pullan1 <sudha.pullalaravu@dell.com> Signed-off-by: Sujit Jadhav <sujit_jadhav@dell.com> Signed-off-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> Signed-off-by: priti-parate <140157516+priti-parate@users.noreply.github.com> Signed-off-by: Kratika Patidar <Kratika.Patidar@dell.com> Co-authored-by: mithileshreddy04 <mithilesh.reddy@dell.com> Co-authored-by: priti-parate <140157516+priti-parate@users.noreply.github.com> Co-authored-by: pullan1 <sudha.pullalaravu@dell.com> Co-authored-by: snarthan <narthan.s@dell.com> Co-authored-by: Sujit Jadhav <sujit_jadhav@dell.com> Co-authored-by: Super User <root@testbed.omnia.test> Co-authored-by: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> Co-authored-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> Co-authored-by: Kratika_Patidar <Kratika.Patidar@dell.com> * IB nic ip assignment * update MinIO and registry images to fixed tagged versions, omnia core container tag and version to 2.2 and v2.2.0.0 (#4309) * Minimal OS-only functional group enablement for x86_64 and aarch64 * Update image_package_collector.py * Update provision_validation.py * Minimal OS functional group updates in provision * Minimal OS functional group upgrade * Fix os_* package cross-contamination and remove stale discovery templates * OpenCHAMI upgrade changes * Update openchami container tags * Update main.yml * Update main.yml * Update main.yml * Update omnia version and core tag * vast client installation * single template way * fix(OMN01D-2164): prompt OME credentials only when enable_bmc_discovery is true - Add enable_bmc_discovery flag (default: false) to discovery_config.yml - Load discovery_config.yml in prepare_oim.yml and set ome_discovery_enabled based on enable_bmc_discovery flag - Change discovery credentials from mandatory to conditional_mandatory gated on ome_discovery_enabled - When enable_bmc_discovery is false, OME username/password prompts are skipped during prepare_oim even if ome_ip is pre-filled * fix(OMN01D-2168): fail explicitly when discovery_mechanism is not provided Replace meta: end_play with ansible.builtin.fail so the playbook exits with non-zero status and a clear error message when discovery_mechanism is missing, instead of silently succeeding. * fix(OMN01D-2169): add L1 input validation for discovery.yml - Create discovery_config.json schema with: - enable_bmc_discovery (boolean, required) - ome_ip (string, required; must be valid IPv4 when enable_bmc_discovery is true) - Register discovery_config in config.py files dict and input_file_inventory - Add 'discovery' tag and invoke validate_config.yml in discovery.yml before role execution (consistent with provision.yml pattern) - Add explicit ome_ip check before OME role inclusion for clear fail-fast error when ome_ip is empty with discovery_mechanism=ome * fix(OMN01D-2225): improve OME authentication and reachability error messages - ome_server_inventory.py: auth failure now tells user to verify ome_username/ome_password in omnia_config_credentials.yml and rerun - collect_inventory.yml: wrap wait_for in block/rescue so timeout gives actionable message pointing to discovery_config.yml and network check * fix(OMN01D-2226): correct discovery completion message next steps - Replace misleading 'Rename or copy' instruction with guidance to update pxe_mapping_file_path in provision_config.yml - Show full absolute path of generated file throughout - Add spacing between steps for readability * fix(OMN01D-2227): escape backslash in docstring to suppress SyntaxWarning Python 3.12+ warns about invalid escape sequence '\d' in non-raw string literals. The docstring in extract_su_from_hostname() contained (?=R\d+) which triggered SyntaxWarning during discovery execution. Escaped the backslash to (?=R\\d+) in the docstring. * fix(OMN01D-2230): correct GROUP_NAME and PARENT_SERVICE_TAG in PXE mapping Issue 1 - GROUP_NAME: - Add fallback: try extracting SU pattern from OME group name when BMC hostname has no SU pattern (covers hierarchical OME groups like SU1_slurm_node, SU2_compute, etc.) - grp0 remains the correct default for single-cluster environments Issue 2 - PARENT_SERVICE_TAG: - Define CHILD_ROLES_OF_CONTROL_PLANE set (service_kube_node_x86_64) - Only assign PARENT_SERVICE_TAG to rows whose FUNCTIONAL_GROUP_NAME is a child role of the control plane within the same GROUP_NAME - Control plane nodes, slurm nodes, login nodes, etc. no longer get an incorrect PARENT_SERVICE_TAG * fix(OMN01D-2231): detect and fail on duplicate OME static group assignments - build_device_group_map() now tracks all group memberships per device and returns a conflicts dict for devices in multiple static groups - main() fails with an actionable error listing each conflicting device and its groups, instead of silently using the first-seen group - Prevents incorrect FUNCTIONAL_GROUP_NAME override in PXE mapping * fix(OMN01D-2232): validate OME group names against supported functional groups - Define SUPPORTED_FUNCTIONAL_GROUPS set matching Omnia's known roles - Skip servers whose OME static group is not in the supported set - Emit a warning per skipped device listing the unsupported group name and the full set of supported groups - Unsupported groups (e.g. 'abc') no longer appear in the PXE mapping * revert: temporarily revert discovery fixes for PR workflow * s3cmd configurations * update image tags * fix: Resolve FK constraint violation and catalog metadata persistence issues - Fix FK constraint by using shared session for image_group_repo and image_repo in ResultPoller - Concatenate multiple S3 paths with semicolon delimiter to respect unique constraint uq_images_image_group_id_role - Update cleanup_job.py to split semicolon-delimited paths and delete each individually - Widen image_name column from VARCHAR(256) to VARCHAR(512) to accommodate semicolon-delimited paths - Add logging to FileArtifactStore.store() to debug file write failures - Fix URI construction in parse_catalog.py for existing artifacts (construct file:// URI directly) - Update build_stream container image tag to 1.1 Tested with job 7da54d1f-ed26-41dd-b3ff-0386104e644d (image-build19) - ImageGroup and Images created successfully * Fix lint error --------- Signed-off-by: balajikumaran.cs <balajikumaran.c.s@gmail.com> Signed-off-by: pullan1 <sudha.pullalaravu@dell.com> Signed-off-by: Sujit Jadhav <sujit_jadhav@dell.com> Signed-off-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> Signed-off-by: priti-parate <140157516+priti-parate@users.noreply.github.com> Signed-off-by: Kratika Patidar <Kratika.Patidar@dell.com> Co-authored-by: Mithilesh Reddy <mithilesh.reddy@dell.com> Co-authored-by: mcas <sakshi.s@dell.com> Co-authored-by: Super User <root@oim.omnia.test> Co-authored-by: snarthan <narthan.s@dell.com> Co-authored-by: balajikumaran.cs <balajikumaran.c.s@gmail.com> Co-authored-by: balajikumaran-c-s <balajikumaran.cs@dellteam.com> Co-authored-by: priti-parate <140157516+priti-parate@users.noreply.github.com> Co-authored-by: Abhishek S A <abhishek.sa3@dell.com> Co-authored-by: pullan1 <sudha.pullalaravu@dell.com> Co-authored-by: Sujit Jadhav <sujit_jadhav@dell.com> Co-authored-by: Super User <root@testbed.omnia.test> Co-authored-by: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> Co-authored-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> Co-authored-by: Kratika_Patidar <Kratika.Patidar@dell.com> Co-authored-by: Nagachandan-P <Nagachandan.p@dell.com> Co-authored-by: Sujit Jadhav <sujit.jadhav@dell.com>

* Minimal OS-only functional group enablement (#4267) * Minimal OS-only functional group enablement for x86_64 and aarch64 * Update image_package_collector.py * Update provision_validation.py * Minimal OS functional group updates in provision * Minimal OS functional group upgrade * Fix os_* package cross-contamination and remove stale discovery templates * csi defect fix * nvidia dcgm install * Add 'provision' tag to omnia_run_tags Signed-off-by: balajikumaran.cs <balajikumaran.c.s@gmail.com> * Add 'provision' tag to omnia_run_tags (#4276) Signed-off-by: balajikumaran.cs <balajikumaran.c.s@gmail.com> * Fix incorrect file path in Podman login failure message (OMN01D-2166) The podman_login_fail_msg referenced a hardcoded path input/omnia_config_credentials.yml which does not exist. Updated it to use the dynamic input_project_dir variable so the error message now correctly points to the actual credentials file at <input_project_dir>/omnia_config_credentials.yml. Fixed in both prepare_oim and gitlab roles. * Merge pull request #4294 from mithileshreddy04/pub/q2_dev OpenCHAMI upgrade changes in prepare_oim and oim_cleanup * Feature branch sync - pub/telemetry to pub/q2_dev (#4293) * Update openchami git version (#4251) Co-authored-by: mithileshreddy04 <mithilesh.reddy@dell.com> Co-authored-by: priti-parate <140157516+priti-parate@users.noreply.github.com> * powerscale teleemtry support with direct authentication mode * use existing vmagent * update messages in vars * merge Pub/q2 dev to pub/telemetry (#4254) * removing input template * Fix for pulp remote RemoteArtifacts is 0 after repo migration Signed-off-by: pullan1 <sudha.pullalaravu@dell.com> --------- Signed-off-by: pullan1 <sudha.pullalaravu@dell.com> Co-authored-by: pullan1 <sudha.pullalaravu@dell.com> Co-authored-by: snarthan <narthan.s@dell.com> * Powerscale teleemtry support using helm * deploy powerscale telemetry using cloud-init * offline deployment of powerscale telemetry * fix for cert-manager failure * fix for cert manager failure * powerscale telemetry deployment with telemetry namespace * sync q2_dev changes (#4263) * removing input template * Fix for pulp remote RemoteArtifacts is 0 after repo migration Signed-off-by: pullan1 <sudha.pullalaravu@dell.com> * Feature/ome discovery pxe mapping enhancements (#4245) * feat(discovery): OME static group extraction, PXE mapping IP/SU/parent tag enhancements ome_server_inventory.py: - Fix static group extraction: find 'Static Groups' container by name and select only direct children via ParentId; avoids picking system/nested groups - Emit module.warn() for static groups that exist but have no devices assigned - Fix idrac_hostname: read InstrumentationName/DnsName from DeviceManagement ManagementType==2 entry instead of DeviceName which returns the IP address generate_pxe_mapping.py: - ADMIN_IP: derive from first 2 octets of admin_network.subnet + last 2 of BMC IP - IB_IP: derive from first 2 octets of ib_network.subnet + last 2 of BMC IP - Skip IB_IP/IB_MAC when server has no IB NIC (ib_nic_mac is empty) - Add extract_su_from_hostname() with regex (SU[A-Z]?\d+)(?=R\d+) to parse Scalable Unit from BMC hostname; rejects service-tag-only hostnames (idrac-JCGT033) and falls back to grp0 when no SU pattern is found - Set GROUP_NAME to extracted SU identifier (fallback: grp0) - Post-process rows to assign PARENT_SERVICE_TAG from the service_kube_control_plane_x86_64 node within the same SU group - Remove BMC_HOSTNAME from CSV headers and output rows - Lint: remove dead try/except in calculate_admin_ip/calculate_ib_ip, reuse ib_mac variable, suppress broad-except pylint warning generate_pxe_mapping.yml: - Load network_spec.yml via include_vars - Set admin_subnet and ib_subnet using selectattr on Networks list - Pass both subnets as parameters to the generate_pxe_mapping module defaults/main.yml: - Add admin_subnet and ib_subnet default variables (empty string) provision_validation.py: - Comment out validate_admin_ips_against_network_spec function and its call site; ADMIN_IPs are now derived from subnet octets + BMC IP and will not necessarily fall within primary_oim_admin_ip/netmask_bits range * refactor: rename discovery directory to provision, update network_spec.yml - Renamed discovery/ to provision/ (git detected as rename, no content loss) - Updated input/network_spec.yml with latest network configuration changes * Update discovery.yml * refactor: unify OME credentials into get_config_credentials flow - Added ome_ip, ome_username, ome_password to omnia_credential.j2 template - Added 'discovery' service entry to omnia_credentials in update_config/vars/main.yml - Added 'discovery' to the hardcoded service key trigger list in fetch_credentials.yml - Replaced custom vault logic in get_ome_credentials.yml with unified decrypt_include_encrypt.yml call against omnia_config_credentials.yml - Updated ome_discovery/vars/main.yml to reference omnia_config_credentials_file and omnia_config_credentials_vault_key instead of the separate .vault/ paths - Deleted .vault/ome_credentials.yml and .vault/.vault_password (no longer needed) * chore: update copyright year from 2025 to 2026 in modified files Updated copyright header in all ome_discovery files modified during this feature branch: - library/generate_pxe_mapping.py - library/ome_server_inventory.py - tasks/generate_pxe_mapping.yml - tasks/get_ome_credentials.yml - defaults/main.yml - vars/main.yml * fix: restore discovery_validations role missed during discovery-to-provision rename discovery/roles/discovery_validations/ was accidentally dropped when renaming the discovery/ directory to provision/. Add it back under provision/roles/discovery_validations/ to resolve the PR merge conflict. * chore: update copyright year to 2026 in provision/roles/discovery_validations files * fix: remove duplicate discovery_validations role (provision_validations already exists) provision/roles/provision_validations/ is the correct renamed equivalent of discovery/roles/discovery_validations/. The discovery_validations copy added to provision/ was redundant. * feat: apply upstream telemetry upgrade changes from dell/omnia pub/q2_dev - Replace kubectl command with kubernetes.core.k8s module for iDRAC StatefulSet - Preserve existing replica count during iDRAC StatefulSet upgrade - Add LDMS store daemon check, restart, and readiness wait tasks * fix: quote build_stream_job_id_absent message in provision_validations vars * feat: add discovery/roles/discovery_validations and telemetry files - Add discovery/roles/discovery_validations/vars/main.yml with task definitions for validation flow - Add discovery/roles/telemetry/tasks/apply_telemetry_on_upgrade.yml with upstream telemetry upgrade logic (replica preservation + LDMS store) * fix: wrap long line in fetch_credentials.yml to satisfy yaml[line-length] lint * refactor: move ome_ip from credentials to discovery_config.yml - Create input/discovery_config.yml for non-credential discovery settings (ome_ip, future Magellan config) - Remove ome_ip from omnia_credential.j2 and credential update vars - Load ome_ip via include_vars from discovery_config.yml in get_ome_credentials.yml - Add discovery_config.yml to provision_validations discovery_inputs - Remove redundant ib_subnet/admin_subnet defaults from ome_discovery * fix: add newline at end of ome_discovery/defaults/main.yml * fix: override role_path to absolute path for decrypt_include_encrypt.yml role_path resolves to ome_discovery role path, causing encrypt_files_vars.yml to be looked up incorrectly. Override to playbook_dir dirname (/opt/omnia/omnia). * fix: inline credential loading to avoid role_path resolution issue role_path cannot be overridden in include_tasks vars. Replace the call to decrypt_include_encrypt.yml with direct include_vars using stat checks for encrypted vs unencrypted credential file handling. * fix: skip load-failure rule in ansible-lint to avoid CI false positives ansible-lint fails to resolve role_path relative paths during static analysis in GitHub Actions, causing false load-failure errors for files that exist and work at runtime. * Update ansible.cfg * Update ansible.cfg * refactor: rename discovery references to provision and add discovery_config variable - Rename discover_mapping_nodes.yml to provision_mapping_nodes.yml - Replace "discovery" terminology with "provision" across playbooks, vars, READMEs, and task names in provision roles - Add subnet as required field with IP pattern validation in network_spec schema - Define discovery_config variable in ome_discovery vars and use it in get_ome_credentials.yml (consistent with provision_config pattern) - Rename discovery_inputs to provision_inputs in validation vars - Rename discovery_mech_mapping to provision_mech_mapping - Update user-facing messages to reference provision.yml Generated with [Devin](https://cli.devin.ai/docs) Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> * fix: credential rules, vault handling, GROUP_NAME validation, and discovery playbook improvements - Add ome_username and ome_password validation rules to credential_rules.json - Add 'discovery' tag to prepare_oim omnia_run_tags so OME credentials are prompted - Fix vault-encrypted credential loading in get_ome_credentials.yml (use decrypt-include-reencrypt pattern instead of unsupported vault_password_file) - Add include_input_dir.yml import to discovery.yml so input_project_dir is set - Accept SU1-SU100 (case-insensitive) in addition to grp0-grp100 for GROUP_NAME - Fix Magellan message to use list format (avoids \n in debug output) - Remove escaped quotes from discovery usage examples Generated with [Devin](https://cli.devin.ai/docs) Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> * fix: extend SU group name support to build_image validation and schemas - Add build_aarch_image tag to input_file_inventory so build_image_aarch64.yml runs provision_config validation (was missing, causing no validation to run for aarch64 builds) - Update GROUP_NAME patterns in functional_groups_config.json and omnia_config.json schemas to accept SU1-SU100 format alongside grp0-grp100 - Update INVALID_GROUP_NAME_MSG to reflect both accepted formats Generated with [Devin](https://cli.devin.ai/docs) Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> --------- Signed-off-by: Sujit Jadhav <sujit_jadhav@dell.com> Co-authored-by: Super User <root@testbed.omnia.test> Co-authored-by: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> * Cleanup discovery roles: move library modules, remove unused roles (#4261) * Cleanup discovery roles: move library modules, remove unused roles - Move ome_server_inventory.py and generate_pxe_mapping.py from discovery/roles/ome_discovery/library/ to common/library/modules/ so they are shared via the common module search path already configured in discovery/ansible.cfg - Remove unused discovery/roles/telemetry/ directory - Remove unused discovery/roles/discovery_validations/ directory - Load discovery_config.yml at playbook level in discovery.yml (consistent with how build_stream_config.yml is loaded in provision.yml) - Fix discovery_complete_msg formatting for readable Ansible output Generated with [Devin](https://cli.devin.ai/docs) Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> * Remove unused discovery_validations role Generated with [Devin](https://cli.devin.ai/docs) Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> --------- Co-authored-by: Super User <root@testbed.omnia.test> Co-authored-by: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> * fix for set_pxe_boot.yml when custom inventory given (#4260) * Update generate_bmc_inventory.yml Signed-off-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> * Update pre_checks.yml Signed-off-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> * lint issue Signed-off-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> --------- Signed-off-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> --------- Signed-off-by: pullan1 <sudha.pullalaravu@dell.com> Signed-off-by: Sujit Jadhav <sujit_jadhav@dell.com> Signed-off-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> Co-authored-by: pullan1 <sudha.pullalaravu@dell.com> Co-authored-by: snarthan <narthan.s@dell.com> Co-authored-by: Sujit Jadhav <sujit_jadhav@dell.com> Co-authored-by: Super User <root@testbed.omnia.test> Co-authored-by: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> Co-authored-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> * resolving merge conflict * revert openchami commit id * resolving review comments * addressing review comments * fix for vmagent scraping powerscale metrics * cleanup script correction for powerscale telemetry cleanup * victoria operator and victoria log input validation * vitoria log input and input validation * remving L2 vslidation for victoria log which is not required * input validation and review comment addressing * change idrac_telemetry_collection_type to telemetry_collection_type * Remove invisible Unicode LRM (U+200E) characters from victoria-operator template filenames * VictoriaLogs container image references and default variable * port check * resolve merge conflict * correction for schema * Update telemetry_config.json * Update validate_input.py * merge conflict telemetry_prereq.yml * change victoria_configurations to victoria_metrics_configurations * remove deployment mode input variable * update for upgrade scenarios * update comments * update comment * resolving issues due to merge conflict * vitoria log changes * victoria log cluster component and VLAgent deployment * updating pod name * removing the changes of adding cert * victoria log changes * remivng victoria log pod calidation playbook * cleanup changes for victoria log * Update ansible-lint.yml and pylint for pub/telemetry (#4296) * Update ansible-lint.yml Signed-off-by: Kratika Patidar <Kratika.Patidar@dell.com> * Update pylint.yml Signed-off-by: Kratika Patidar <Kratika.Patidar@dell.com> * fixing ansible-lint * lint * line-lenght --------- Signed-off-by: Kratika Patidar <Kratika.Patidar@dell.com> --------- Signed-off-by: pullan1 <sudha.pullalaravu@dell.com> Signed-off-by: Sujit Jadhav <sujit_jadhav@dell.com> Signed-off-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> Signed-off-by: priti-parate <140157516+priti-parate@users.noreply.github.com> Signed-off-by: Kratika Patidar <Kratika.Patidar@dell.com> Co-authored-by: mithileshreddy04 <mithilesh.reddy@dell.com> Co-authored-by: priti-parate <140157516+priti-parate@users.noreply.github.com> Co-authored-by: pullan1 <sudha.pullalaravu@dell.com> Co-authored-by: snarthan <narthan.s@dell.com> Co-authored-by: Sujit Jadhav <sujit_jadhav@dell.com> Co-authored-by: Super User <root@testbed.omnia.test> Co-authored-by: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> Co-authored-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> Co-authored-by: Kratika_Patidar <Kratika.Patidar@dell.com> * IB nic ip assignment * update MinIO and registry images to fixed tagged versions, omnia core container tag and version to 2.2 and v2.2.0.0 (#4309) * Minimal OS-only functional group enablement for x86_64 and aarch64 * Update image_package_collector.py * Update provision_validation.py * Minimal OS functional group updates in provision * Minimal OS functional group upgrade * Fix os_* package cross-contamination and remove stale discovery templates * OpenCHAMI upgrade changes * Update openchami container tags * Update main.yml * Update main.yml * Update main.yml * Update omnia version and core tag * vast client installation * single template way * fix(OMN01D-2164): prompt OME credentials only when enable_bmc_discovery is true - Add enable_bmc_discovery flag (default: false) to discovery_config.yml - Load discovery_config.yml in prepare_oim.yml and set ome_discovery_enabled based on enable_bmc_discovery flag - Change discovery credentials from mandatory to conditional_mandatory gated on ome_discovery_enabled - When enable_bmc_discovery is false, OME username/password prompts are skipped during prepare_oim even if ome_ip is pre-filled * fix(OMN01D-2168): fail explicitly when discovery_mechanism is not provided Replace meta: end_play with ansible.builtin.fail so the playbook exits with non-zero status and a clear error message when discovery_mechanism is missing, instead of silently succeeding. * fix(OMN01D-2169): add L1 input validation for discovery.yml - Create discovery_config.json schema with: - enable_bmc_discovery (boolean, required) - ome_ip (string, required; must be valid IPv4 when enable_bmc_discovery is true) - Register discovery_config in config.py files dict and input_file_inventory - Add 'discovery' tag and invoke validate_config.yml in discovery.yml before role execution (consistent with provision.yml pattern) - Add explicit ome_ip check before OME role inclusion for clear fail-fast error when ome_ip is empty with discovery_mechanism=ome * fix(OMN01D-2225): improve OME authentication and reachability error messages - ome_server_inventory.py: auth failure now tells user to verify ome_username/ome_password in omnia_config_credentials.yml and rerun - collect_inventory.yml: wrap wait_for in block/rescue so timeout gives actionable message pointing to discovery_config.yml and network check * fix(OMN01D-2226): correct discovery completion message next steps - Replace misleading 'Rename or copy' instruction with guidance to update pxe_mapping_file_path in provision_config.yml - Show full absolute path of generated file throughout - Add spacing between steps for readability * fix(OMN01D-2227): escape backslash in docstring to suppress SyntaxWarning Python 3.12+ warns about invalid escape sequence '\d' in non-raw string literals. The docstring in extract_su_from_hostname() contained (?=R\d+) which triggered SyntaxWarning during discovery execution. Escaped the backslash to (?=R\\d+) in the docstring. * fix(OMN01D-2230): correct GROUP_NAME and PARENT_SERVICE_TAG in PXE mapping Issue 1 - GROUP_NAME: - Add fallback: try extracting SU pattern from OME group name when BMC hostname has no SU pattern (covers hierarchical OME groups like SU1_slurm_node, SU2_compute, etc.) - grp0 remains the correct default for single-cluster environments Issue 2 - PARENT_SERVICE_TAG: - Define CHILD_ROLES_OF_CONTROL_PLANE set (service_kube_node_x86_64) - Only assign PARENT_SERVICE_TAG to rows whose FUNCTIONAL_GROUP_NAME is a child role of the control plane within the same GROUP_NAME - Control plane nodes, slurm nodes, login nodes, etc. no longer get an incorrect PARENT_SERVICE_TAG * fix(OMN01D-2231): detect and fail on duplicate OME static group assignments - build_device_group_map() now tracks all group memberships per device and returns a conflicts dict for devices in multiple static groups - main() fails with an actionable error listing each conflicting device and its groups, instead of silently using the first-seen group - Prevents incorrect FUNCTIONAL_GROUP_NAME override in PXE mapping * fix(OMN01D-2232): validate OME group names against supported functional groups - Define SUPPORTED_FUNCTIONAL_GROUPS set matching Omnia's known roles - Skip servers whose OME static group is not in the supported set - Emit a warning per skipped device listing the unsupported group name and the full set of supported groups - Unsupported groups (e.g. 'abc') no longer appear in the PXE mapping * revert: temporarily revert discovery fixes for PR workflow * s3cmd configurations * update image tags * fix: Resolve FK constraint violation and catalog metadata persistence issues - Fix FK constraint by using shared session for image_group_repo and image_repo in ResultPoller - Concatenate multiple S3 paths with semicolon delimiter to respect unique constraint uq_images_image_group_id_role - Update cleanup_job.py to split semicolon-delimited paths and delete each individually - Widen image_name column from VARCHAR(256) to VARCHAR(512) to accommodate semicolon-delimited paths - Add logging to FileArtifactStore.store() to debug file write failures - Fix URI construction in parse_catalog.py for existing artifacts (construct file:// URI directly) - Update build_stream container image tag to 1.1 Tested with job 7da54d1f-ed26-41dd-b3ff-0386104e644d (image-build19) - ImageGroup and Images created successfully * Fix lint error * Modify the charlimit to 512 in DB --------- Signed-off-by: balajikumaran.cs <balajikumaran.c.s@gmail.com> Signed-off-by: pullan1 <sudha.pullalaravu@dell.com> Signed-off-by: Sujit Jadhav <sujit_jadhav@dell.com> Signed-off-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> Signed-off-by: priti-parate <140157516+priti-parate@users.noreply.github.com> Signed-off-by: Kratika Patidar <Kratika.Patidar@dell.com> Co-authored-by: Mithilesh Reddy <mithilesh.reddy@dell.com> Co-authored-by: mcas <sakshi.s@dell.com> Co-authored-by: Super User <root@oim.omnia.test> Co-authored-by: snarthan <narthan.s@dell.com> Co-authored-by: balajikumaran.cs <balajikumaran.c.s@gmail.com> Co-authored-by: balajikumaran-c-s <balajikumaran.cs@dellteam.com> Co-authored-by: priti-parate <140157516+priti-parate@users.noreply.github.com> Co-authored-by: Abhishek S A <abhishek.sa3@dell.com> Co-authored-by: pullan1 <sudha.pullalaravu@dell.com> Co-authored-by: Sujit Jadhav <sujit_jadhav@dell.com> Co-authored-by: Super User <root@testbed.omnia.test> Co-authored-by: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> Co-authored-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> Co-authored-by: Kratika_Patidar <Kratika.Patidar@dell.com> Co-authored-by: Nagachandan-P <Nagachandan.p@dell.com> Co-authored-by: Sujit Jadhav <sujit.jadhav@dell.com>

* Minimal OS-only functional group enablement (#4267) * Minimal OS-only functional group enablement for x86_64 and aarch64 * Update image_package_collector.py * Update provision_validation.py * Minimal OS functional group updates in provision * Minimal OS functional group upgrade * Fix os_* package cross-contamination and remove stale discovery templates * csi defect fix * nvidia dcgm install * Add 'provision' tag to omnia_run_tags Signed-off-by: balajikumaran.cs <balajikumaran.c.s@gmail.com> * Add 'provision' tag to omnia_run_tags (#4276) Signed-off-by: balajikumaran.cs <balajikumaran.c.s@gmail.com> * Fix incorrect file path in Podman login failure message (OMN01D-2166) The podman_login_fail_msg referenced a hardcoded path input/omnia_config_credentials.yml which does not exist. Updated it to use the dynamic input_project_dir variable so the error message now correctly points to the actual credentials file at <input_project_dir>/omnia_config_credentials.yml. Fixed in both prepare_oim and gitlab roles. * Merge pull request #4294 from mithileshreddy04/pub/q2_dev OpenCHAMI upgrade changes in prepare_oim and oim_cleanup * Feature branch sync - pub/telemetry to pub/q2_dev (#4293) * Update openchami git version (#4251) Co-authored-by: mithileshreddy04 <mithilesh.reddy@dell.com> Co-authored-by: priti-parate <140157516+priti-parate@users.noreply.github.com> * powerscale teleemtry support with direct authentication mode * use existing vmagent * update messages in vars * merge Pub/q2 dev to pub/telemetry (#4254) * removing input template * Fix for pulp remote RemoteArtifacts is 0 after repo migration Signed-off-by: pullan1 <sudha.pullalaravu@dell.com> --------- Signed-off-by: pullan1 <sudha.pullalaravu@dell.com> Co-authored-by: pullan1 <sudha.pullalaravu@dell.com> Co-authored-by: snarthan <narthan.s@dell.com> * Powerscale teleemtry support using helm * deploy powerscale telemetry using cloud-init * offline deployment of powerscale telemetry * fix for cert-manager failure * fix for cert manager failure * powerscale telemetry deployment with telemetry namespace * sync q2_dev changes (#4263) * removing input template * Fix for pulp remote RemoteArtifacts is 0 after repo migration Signed-off-by: pullan1 <sudha.pullalaravu@dell.com> * Feature/ome discovery pxe mapping enhancements (#4245) * feat(discovery): OME static group extraction, PXE mapping IP/SU/parent tag enhancements ome_server_inventory.py: - Fix static group extraction: find 'Static Groups' container by name and select only direct children via ParentId; avoids picking system/nested groups - Emit module.warn() for static groups that exist but have no devices assigned - Fix idrac_hostname: read InstrumentationName/DnsName from DeviceManagement ManagementType==2 entry instead of DeviceName which returns the IP address generate_pxe_mapping.py: - ADMIN_IP: derive from first 2 octets of admin_network.subnet + last 2 of BMC IP - IB_IP: derive from first 2 octets of ib_network.subnet + last 2 of BMC IP - Skip IB_IP/IB_MAC when server has no IB NIC (ib_nic_mac is empty) - Add extract_su_from_hostname() with regex (SU[A-Z]?\d+)(?=R\d+) to parse Scalable Unit from BMC hostname; rejects service-tag-only hostnames (idrac-JCGT033) and falls back to grp0 when no SU pattern is found - Set GROUP_NAME to extracted SU identifier (fallback: grp0) - Post-process rows to assign PARENT_SERVICE_TAG from the service_kube_control_plane_x86_64 node within the same SU group - Remove BMC_HOSTNAME from CSV headers and output rows - Lint: remove dead try/except in calculate_admin_ip/calculate_ib_ip, reuse ib_mac variable, suppress broad-except pylint warning generate_pxe_mapping.yml: - Load network_spec.yml via include_vars - Set admin_subnet and ib_subnet using selectattr on Networks list - Pass both subnets as parameters to the generate_pxe_mapping module defaults/main.yml: - Add admin_subnet and ib_subnet default variables (empty string) provision_validation.py: - Comment out validate_admin_ips_against_network_spec function and its call site; ADMIN_IPs are now derived from subnet octets + BMC IP and will not necessarily fall within primary_oim_admin_ip/netmask_bits range * refactor: rename discovery directory to provision, update network_spec.yml - Renamed discovery/ to provision/ (git detected as rename, no content loss) - Updated input/network_spec.yml with latest network configuration changes * Update discovery.yml * refactor: unify OME credentials into get_config_credentials flow - Added ome_ip, ome_username, ome_password to omnia_credential.j2 template - Added 'discovery' service entry to omnia_credentials in update_config/vars/main.yml - Added 'discovery' to the hardcoded service key trigger list in fetch_credentials.yml - Replaced custom vault logic in get_ome_credentials.yml with unified decrypt_include_encrypt.yml call against omnia_config_credentials.yml - Updated ome_discovery/vars/main.yml to reference omnia_config_credentials_file and omnia_config_credentials_vault_key instead of the separate .vault/ paths - Deleted .vault/ome_credentials.yml and .vault/.vault_password (no longer needed) * chore: update copyright year from 2025 to 2026 in modified files Updated copyright header in all ome_discovery files modified during this feature branch: - library/generate_pxe_mapping.py - library/ome_server_inventory.py - tasks/generate_pxe_mapping.yml - tasks/get_ome_credentials.yml - defaults/main.yml - vars/main.yml * fix: restore discovery_validations role missed during discovery-to-provision rename discovery/roles/discovery_validations/ was accidentally dropped when renaming the discovery/ directory to provision/. Add it back under provision/roles/discovery_validations/ to resolve the PR merge conflict. * chore: update copyright year to 2026 in provision/roles/discovery_validations files * fix: remove duplicate discovery_validations role (provision_validations already exists) provision/roles/provision_validations/ is the correct renamed equivalent of discovery/roles/discovery_validations/. The discovery_validations copy added to provision/ was redundant. * feat: apply upstream telemetry upgrade changes from dell/omnia pub/q2_dev - Replace kubectl command with kubernetes.core.k8s module for iDRAC StatefulSet - Preserve existing replica count during iDRAC StatefulSet upgrade - Add LDMS store daemon check, restart, and readiness wait tasks * fix: quote build_stream_job_id_absent message in provision_validations vars * feat: add discovery/roles/discovery_validations and telemetry files - Add discovery/roles/discovery_validations/vars/main.yml with task definitions for validation flow - Add discovery/roles/telemetry/tasks/apply_telemetry_on_upgrade.yml with upstream telemetry upgrade logic (replica preservation + LDMS store) * fix: wrap long line in fetch_credentials.yml to satisfy yaml[line-length] lint * refactor: move ome_ip from credentials to discovery_config.yml - Create input/discovery_config.yml for non-credential discovery settings (ome_ip, future Magellan config) - Remove ome_ip from omnia_credential.j2 and credential update vars - Load ome_ip via include_vars from discovery_config.yml in get_ome_credentials.yml - Add discovery_config.yml to provision_validations discovery_inputs - Remove redundant ib_subnet/admin_subnet defaults from ome_discovery * fix: add newline at end of ome_discovery/defaults/main.yml * fix: override role_path to absolute path for decrypt_include_encrypt.yml role_path resolves to ome_discovery role path, causing encrypt_files_vars.yml to be looked up incorrectly. Override to playbook_dir dirname (/opt/omnia/omnia). * fix: inline credential loading to avoid role_path resolution issue role_path cannot be overridden in include_tasks vars. Replace the call to decrypt_include_encrypt.yml with direct include_vars using stat checks for encrypted vs unencrypted credential file handling. * fix: skip load-failure rule in ansible-lint to avoid CI false positives ansible-lint fails to resolve role_path relative paths during static analysis in GitHub Actions, causing false load-failure errors for files that exist and work at runtime. * Update ansible.cfg * Update ansible.cfg * refactor: rename discovery references to provision and add discovery_config variable - Rename discover_mapping_nodes.yml to provision_mapping_nodes.yml - Replace "discovery" terminology with "provision" across playbooks, vars, READMEs, and task names in provision roles - Add subnet as required field with IP pattern validation in network_spec schema - Define discovery_config variable in ome_discovery vars and use it in get_ome_credentials.yml (consistent with provision_config pattern) - Rename discovery_inputs to provision_inputs in validation vars - Rename discovery_mech_mapping to provision_mech_mapping - Update user-facing messages to reference provision.yml Generated with [Devin](https://cli.devin.ai/docs) Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> * fix: credential rules, vault handling, GROUP_NAME validation, and discovery playbook improvements - Add ome_username and ome_password validation rules to credential_rules.json - Add 'discovery' tag to prepare_oim omnia_run_tags so OME credentials are prompted - Fix vault-encrypted credential loading in get_ome_credentials.yml (use decrypt-include-reencrypt pattern instead of unsupported vault_password_file) - Add include_input_dir.yml import to discovery.yml so input_project_dir is set - Accept SU1-SU100 (case-insensitive) in addition to grp0-grp100 for GROUP_NAME - Fix Magellan message to use list format (avoids \n in debug output) - Remove escaped quotes from discovery usage examples Generated with [Devin](https://cli.devin.ai/docs) Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> * fix: extend SU group name support to build_image validation and schemas - Add build_aarch_image tag to input_file_inventory so build_image_aarch64.yml runs provision_config validation (was missing, causing no validation to run for aarch64 builds) - Update GROUP_NAME patterns in functional_groups_config.json and omnia_config.json schemas to accept SU1-SU100 format alongside grp0-grp100 - Update INVALID_GROUP_NAME_MSG to reflect both accepted formats Generated with [Devin](https://cli.devin.ai/docs) Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> --------- Signed-off-by: Sujit Jadhav <sujit_jadhav@dell.com> Co-authored-by: Super User <root@testbed.omnia.test> Co-authored-by: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> * Cleanup discovery roles: move library modules, remove unused roles (#4261) * Cleanup discovery roles: move library modules, remove unused roles - Move ome_server_inventory.py and generate_pxe_mapping.py from discovery/roles/ome_discovery/library/ to common/library/modules/ so they are shared via the common module search path already configured in discovery/ansible.cfg - Remove unused discovery/roles/telemetry/ directory - Remove unused discovery/roles/discovery_validations/ directory - Load discovery_config.yml at playbook level in discovery.yml (consistent with how build_stream_config.yml is loaded in provision.yml) - Fix discovery_complete_msg formatting for readable Ansible output Generated with [Devin](https://cli.devin.ai/docs) Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> * Remove unused discovery_validations role Generated with [Devin](https://cli.devin.ai/docs) Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> --------- Co-authored-by: Super User <root@testbed.omnia.test> Co-authored-by: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> * fix for set_pxe_boot.yml when custom inventory given (#4260) * Update generate_bmc_inventory.yml Signed-off-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> * Update pre_checks.yml Signed-off-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> * lint issue Signed-off-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> --------- Signed-off-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> --------- Signed-off-by: pullan1 <sudha.pullalaravu@dell.com> Signed-off-by: Sujit Jadhav <sujit_jadhav@dell.com> Signed-off-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> Co-authored-by: pullan1 <sudha.pullalaravu@dell.com> Co-authored-by: snarthan <narthan.s@dell.com> Co-authored-by: Sujit Jadhav <sujit_jadhav@dell.com> Co-authored-by: Super User <root@testbed.omnia.test> Co-authored-by: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> Co-authored-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> * resolving merge conflict * revert openchami commit id * resolving review comments * addressing review comments * fix for vmagent scraping powerscale metrics * cleanup script correction for powerscale telemetry cleanup * victoria operator and victoria log input validation * vitoria log input and input validation * remving L2 vslidation for victoria log which is not required * input validation and review comment addressing * change idrac_telemetry_collection_type to telemetry_collection_type * Remove invisible Unicode LRM (U+200E) characters from victoria-operator template filenames * VictoriaLogs container image references and default variable * port check * resolve merge conflict * correction for schema * Update telemetry_config.json * Update validate_input.py * merge conflict telemetry_prereq.yml * change victoria_configurations to victoria_metrics_configurations * remove deployment mode input variable * update for upgrade scenarios * update comments * update comment * resolving issues due to merge conflict * vitoria log changes * victoria log cluster component and VLAgent deployment * updating pod name * removing the changes of adding cert * victoria log changes * remivng victoria log pod calidation playbook * cleanup changes for victoria log * Update ansible-lint.yml and pylint for pub/telemetry (#4296) * Update ansible-lint.yml Signed-off-by: Kratika Patidar <Kratika.Patidar@dell.com> * Update pylint.yml Signed-off-by: Kratika Patidar <Kratika.Patidar@dell.com> * fixing ansible-lint * lint * line-lenght --------- Signed-off-by: Kratika Patidar <Kratika.Patidar@dell.com> --------- Signed-off-by: pullan1 <sudha.pullalaravu@dell.com> Signed-off-by: Sujit Jadhav <sujit_jadhav@dell.com> Signed-off-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> Signed-off-by: priti-parate <140157516+priti-parate@users.noreply.github.com> Signed-off-by: Kratika Patidar <Kratika.Patidar@dell.com> Co-authored-by: mithileshreddy04 <mithilesh.reddy@dell.com> Co-authored-by: priti-parate <140157516+priti-parate@users.noreply.github.com> Co-authored-by: pullan1 <sudha.pullalaravu@dell.com> Co-authored-by: snarthan <narthan.s@dell.com> Co-authored-by: Sujit Jadhav <sujit_jadhav@dell.com> Co-authored-by: Super User <root@testbed.omnia.test> Co-authored-by: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> Co-authored-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> Co-authored-by: Kratika_Patidar <Kratika.Patidar@dell.com> * IB nic ip assignment * update MinIO and registry images to fixed tagged versions, omnia core container tag and version to 2.2 and v2.2.0.0 (#4309) * Minimal OS-only functional group enablement for x86_64 and aarch64 * Update image_package_collector.py * Update provision_validation.py * Minimal OS functional group updates in provision * Minimal OS functional group upgrade * Fix os_* package cross-contamination and remove stale discovery templates * OpenCHAMI upgrade changes * Update openchami container tags * Update main.yml * Update main.yml * Update main.yml * Update omnia version and core tag * vast client installation * single template way * fix(OMN01D-2164): prompt OME credentials only when enable_bmc_discovery is true - Add enable_bmc_discovery flag (default: false) to discovery_config.yml - Load discovery_config.yml in prepare_oim.yml and set ome_discovery_enabled based on enable_bmc_discovery flag - Change discovery credentials from mandatory to conditional_mandatory gated on ome_discovery_enabled - When enable_bmc_discovery is false, OME username/password prompts are skipped during prepare_oim even if ome_ip is pre-filled * fix(OMN01D-2168): fail explicitly when discovery_mechanism is not provided Replace meta: end_play with ansible.builtin.fail so the playbook exits with non-zero status and a clear error message when discovery_mechanism is missing, instead of silently succeeding. * fix(OMN01D-2169): add L1 input validation for discovery.yml - Create discovery_config.json schema with: - enable_bmc_discovery (boolean, required) - ome_ip (string, required; must be valid IPv4 when enable_bmc_discovery is true) - Register discovery_config in config.py files dict and input_file_inventory - Add 'discovery' tag and invoke validate_config.yml in discovery.yml before role execution (consistent with provision.yml pattern) - Add explicit ome_ip check before OME role inclusion for clear fail-fast error when ome_ip is empty with discovery_mechanism=ome * fix(OMN01D-2225): improve OME authentication and reachability error messages - ome_server_inventory.py: auth failure now tells user to verify ome_username/ome_password in omnia_config_credentials.yml and rerun - collect_inventory.yml: wrap wait_for in block/rescue so timeout gives actionable message pointing to discovery_config.yml and network check * fix(OMN01D-2226): correct discovery completion message next steps - Replace misleading 'Rename or copy' instruction with guidance to update pxe_mapping_file_path in provision_config.yml - Show full absolute path of generated file throughout - Add spacing between steps for readability * fix(OMN01D-2227): escape backslash in docstring to suppress SyntaxWarning Python 3.12+ warns about invalid escape sequence '\d' in non-raw string literals. The docstring in extract_su_from_hostname() contained (?=R\d+) which triggered SyntaxWarning during discovery execution. Escaped the backslash to (?=R\\d+) in the docstring. * fix(OMN01D-2230): correct GROUP_NAME and PARENT_SERVICE_TAG in PXE mapping Issue 1 - GROUP_NAME: - Add fallback: try extracting SU pattern from OME group name when BMC hostname has no SU pattern (covers hierarchical OME groups like SU1_slurm_node, SU2_compute, etc.) - grp0 remains the correct default for single-cluster environments Issue 2 - PARENT_SERVICE_TAG: - Define CHILD_ROLES_OF_CONTROL_PLANE set (service_kube_node_x86_64) - Only assign PARENT_SERVICE_TAG to rows whose FUNCTIONAL_GROUP_NAME is a child role of the control plane within the same GROUP_NAME - Control plane nodes, slurm nodes, login nodes, etc. no longer get an incorrect PARENT_SERVICE_TAG * fix(OMN01D-2231): detect and fail on duplicate OME static group assignments - build_device_group_map() now tracks all group memberships per device and returns a conflicts dict for devices in multiple static groups - main() fails with an actionable error listing each conflicting device and its groups, instead of silently using the first-seen group - Prevents incorrect FUNCTIONAL_GROUP_NAME override in PXE mapping * fix(OMN01D-2232): validate OME group names against supported functional groups - Define SUPPORTED_FUNCTIONAL_GROUPS set matching Omnia's known roles - Skip servers whose OME static group is not in the supported set - Emit a warning per skipped device listing the unsupported group name and the full set of supported groups - Unsupported groups (e.g. 'abc') no longer appear in the PXE mapping * revert: temporarily revert discovery fixes for PR workflow * s3cmd configurations * update image tags * fix: Resolve FK constraint violation and catalog metadata persistence issues - Fix FK constraint by using shared session for image_group_repo and image_repo in ResultPoller - Concatenate multiple S3 paths with semicolon delimiter to respect unique constraint uq_images_image_group_id_role - Update cleanup_job.py to split semicolon-delimited paths and delete each individually - Widen image_name column from VARCHAR(256) to VARCHAR(512) to accommodate semicolon-delimited paths - Add logging to FileArtifactStore.store() to debug file write failures - Fix URI construction in parse_catalog.py for existing artifacts (construct file:// URI directly) - Update build_stream container image tag to 1.1 Tested with job 7da54d1f-ed26-41dd-b3ff-0386104e644d (image-build19) - ImageGroup and Images created successfully * Fix lint error * Modify the charlimit to 512 in DB * Fix the image_groups name and stages in summary * Updating Catalog changes aligning to recent updates * Fix the pylint issues --------- Signed-off-by: balajikumaran.cs <balajikumaran.c.s@gmail.com> Signed-off-by: pullan1 <sudha.pullalaravu@dell.com> Signed-off-by: Sujit Jadhav <sujit_jadhav@dell.com> Signed-off-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> Signed-off-by: priti-parate <140157516+priti-parate@users.noreply.github.com> Signed-off-by: Kratika Patidar <Kratika.Patidar@dell.com> Co-authored-by: Mithilesh Reddy <mithilesh.reddy@dell.com> Co-authored-by: mcas <sakshi.s@dell.com> Co-authored-by: Super User <root@oim.omnia.test> Co-authored-by: snarthan <narthan.s@dell.com> Co-authored-by: balajikumaran.cs <balajikumaran.c.s@gmail.com> Co-authored-by: balajikumaran-c-s <balajikumaran.cs@dellteam.com> Co-authored-by: priti-parate <140157516+priti-parate@users.noreply.github.com> Co-authored-by: Abhishek S A <abhishek.sa3@dell.com> Co-authored-by: pullan1 <sudha.pullalaravu@dell.com> Co-authored-by: Sujit Jadhav <sujit_jadhav@dell.com> Co-authored-by: Super User <root@testbed.omnia.test> Co-authored-by: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> Co-authored-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> Co-authored-by: Kratika_Patidar <Kratika.Patidar@dell.com> Co-authored-by: Nagachandan-P <Nagachandan.p@dell.com> Co-authored-by: Sujit Jadhav <sujit.jadhav@dell.com>

* Minimal OS-only functional group enablement (#4267) * Minimal OS-only functional group enablement for x86_64 and aarch64 * Update image_package_collector.py * Update provision_validation.py * Minimal OS functional group updates in provision * Minimal OS functional group upgrade * Fix os_* package cross-contamination and remove stale discovery templates * csi defect fix * nvidia dcgm install * Add 'provision' tag to omnia_run_tags Signed-off-by: balajikumaran.cs <balajikumaran.c.s@gmail.com> * Add 'provision' tag to omnia_run_tags (#4276) Signed-off-by: balajikumaran.cs <balajikumaran.c.s@gmail.com> * Fix incorrect file path in Podman login failure message (OMN01D-2166) The podman_login_fail_msg referenced a hardcoded path input/omnia_config_credentials.yml which does not exist. Updated it to use the dynamic input_project_dir variable so the error message now correctly points to the actual credentials file at <input_project_dir>/omnia_config_credentials.yml. Fixed in both prepare_oim and gitlab roles. * Merge pull request #4294 from mithileshreddy04/pub/q2_dev OpenCHAMI upgrade changes in prepare_oim and oim_cleanup * Feature branch sync - pub/telemetry to pub/q2_dev (#4293) * Update openchami git version (#4251) Co-authored-by: mithileshreddy04 <mithilesh.reddy@dell.com> Co-authored-by: priti-parate <140157516+priti-parate@users.noreply.github.com> * powerscale teleemtry support with direct authentication mode * use existing vmagent * update messages in vars * merge Pub/q2 dev to pub/telemetry (#4254) * removing input template * Fix for pulp remote RemoteArtifacts is 0 after repo migration Signed-off-by: pullan1 <sudha.pullalaravu@dell.com> --------- Signed-off-by: pullan1 <sudha.pullalaravu@dell.com> Co-authored-by: pullan1 <sudha.pullalaravu@dell.com> Co-authored-by: snarthan <narthan.s@dell.com> * Powerscale teleemtry support using helm * deploy powerscale telemetry using cloud-init * offline deployment of powerscale telemetry * fix for cert-manager failure * fix for cert manager failure * powerscale telemetry deployment with telemetry namespace * sync q2_dev changes (#4263) * removing input template * Fix for pulp remote RemoteArtifacts is 0 after repo migration Signed-off-by: pullan1 <sudha.pullalaravu@dell.com> * Feature/ome discovery pxe mapping enhancements (#4245) * feat(discovery): OME static group extraction, PXE mapping IP/SU/parent tag enhancements ome_server_inventory.py: - Fix static group extraction: find 'Static Groups' container by name and select only direct children via ParentId; avoids picking system/nested groups - Emit module.warn() for static groups that exist but have no devices assigned - Fix idrac_hostname: read InstrumentationName/DnsName from DeviceManagement ManagementType==2 entry instead of DeviceName which returns the IP address generate_pxe_mapping.py: - ADMIN_IP: derive from first 2 octets of admin_network.subnet + last 2 of BMC IP - IB_IP: derive from first 2 octets of ib_network.subnet + last 2 of BMC IP - Skip IB_IP/IB_MAC when server has no IB NIC (ib_nic_mac is empty) - Add extract_su_from_hostname() with regex (SU[A-Z]?\d+)(?=R\d+) to parse Scalable Unit from BMC hostname; rejects service-tag-only hostnames (idrac-JCGT033) and falls back to grp0 when no SU pattern is found - Set GROUP_NAME to extracted SU identifier (fallback: grp0) - Post-process rows to assign PARENT_SERVICE_TAG from the service_kube_control_plane_x86_64 node within the same SU group - Remove BMC_HOSTNAME from CSV headers and output rows - Lint: remove dead try/except in calculate_admin_ip/calculate_ib_ip, reuse ib_mac variable, suppress broad-except pylint warning generate_pxe_mapping.yml: - Load network_spec.yml via include_vars - Set admin_subnet and ib_subnet using selectattr on Networks list - Pass both subnets as parameters to the generate_pxe_mapping module defaults/main.yml: - Add admin_subnet and ib_subnet default variables (empty string) provision_validation.py: - Comment out validate_admin_ips_against_network_spec function and its call site; ADMIN_IPs are now derived from subnet octets + BMC IP and will not necessarily fall within primary_oim_admin_ip/netmask_bits range * refactor: rename discovery directory to provision, update network_spec.yml - Renamed discovery/ to provision/ (git detected as rename, no content loss) - Updated input/network_spec.yml with latest network configuration changes * Update discovery.yml * refactor: unify OME credentials into get_config_credentials flow - Added ome_ip, ome_username, ome_password to omnia_credential.j2 template - Added 'discovery' service entry to omnia_credentials in update_config/vars/main.yml - Added 'discovery' to the hardcoded service key trigger list in fetch_credentials.yml - Replaced custom vault logic in get_ome_credentials.yml with unified decrypt_include_encrypt.yml call against omnia_config_credentials.yml - Updated ome_discovery/vars/main.yml to reference omnia_config_credentials_file and omnia_config_credentials_vault_key instead of the separate .vault/ paths - Deleted .vault/ome_credentials.yml and .vault/.vault_password (no longer needed) * chore: update copyright year from 2025 to 2026 in modified files Updated copyright header in all ome_discovery files modified during this feature branch: - library/generate_pxe_mapping.py - library/ome_server_inventory.py - tasks/generate_pxe_mapping.yml - tasks/get_ome_credentials.yml - defaults/main.yml - vars/main.yml * fix: restore discovery_validations role missed during discovery-to-provision rename discovery/roles/discovery_validations/ was accidentally dropped when renaming the discovery/ directory to provision/. Add it back under provision/roles/discovery_validations/ to resolve the PR merge conflict. * chore: update copyright year to 2026 in provision/roles/discovery_validations files * fix: remove duplicate discovery_validations role (provision_validations already exists) provision/roles/provision_validations/ is the correct renamed equivalent of discovery/roles/discovery_validations/. The discovery_validations copy added to provision/ was redundant. * feat: apply upstream telemetry upgrade changes from dell/omnia pub/q2_dev - Replace kubectl command with kubernetes.core.k8s module for iDRAC StatefulSet - Preserve existing replica count during iDRAC StatefulSet upgrade - Add LDMS store daemon check, restart, and readiness wait tasks * fix: quote build_stream_job_id_absent message in provision_validations vars * feat: add discovery/roles/discovery_validations and telemetry files - Add discovery/roles/discovery_validations/vars/main.yml with task definitions for validation flow - Add discovery/roles/telemetry/tasks/apply_telemetry_on_upgrade.yml with upstream telemetry upgrade logic (replica preservation + LDMS store) * fix: wrap long line in fetch_credentials.yml to satisfy yaml[line-length] lint * refactor: move ome_ip from credentials to discovery_config.yml - Create input/discovery_config.yml for non-credential discovery settings (ome_ip, future Magellan config) - Remove ome_ip from omnia_credential.j2 and credential update vars - Load ome_ip via include_vars from discovery_config.yml in get_ome_credentials.yml - Add discovery_config.yml to provision_validations discovery_inputs - Remove redundant ib_subnet/admin_subnet defaults from ome_discovery * fix: add newline at end of ome_discovery/defaults/main.yml * fix: override role_path to absolute path for decrypt_include_encrypt.yml role_path resolves to ome_discovery role path, causing encrypt_files_vars.yml to be looked up incorrectly. Override to playbook_dir dirname (/opt/omnia/omnia). * fix: inline credential loading to avoid role_path resolution issue role_path cannot be overridden in include_tasks vars. Replace the call to decrypt_include_encrypt.yml with direct include_vars using stat checks for encrypted vs unencrypted credential file handling. * fix: skip load-failure rule in ansible-lint to avoid CI false positives ansible-lint fails to resolve role_path relative paths during static analysis in GitHub Actions, causing false load-failure errors for files that exist and work at runtime. * Update ansible.cfg * Update ansible.cfg * refactor: rename discovery references to provision and add discovery_config variable - Rename discover_mapping_nodes.yml to provision_mapping_nodes.yml - Replace "discovery" terminology with "provision" across playbooks, vars, READMEs, and task names in provision roles - Add subnet as required field with IP pattern validation in network_spec schema - Define discovery_config variable in ome_discovery vars and use it in get_ome_credentials.yml (consistent with provision_config pattern) - Rename discovery_inputs to provision_inputs in validation vars - Rename discovery_mech_mapping to provision_mech_mapping - Update user-facing messages to reference provision.yml Generated with [Devin](https://cli.devin.ai/docs) Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> * fix: credential rules, vault handling, GROUP_NAME validation, and discovery playbook improvements - Add ome_username and ome_password validation rules to credential_rules.json - Add 'discovery' tag to prepare_oim omnia_run_tags so OME credentials are prompted - Fix vault-encrypted credential loading in get_ome_credentials.yml (use decrypt-include-reencrypt pattern instead of unsupported vault_password_file) - Add include_input_dir.yml import to discovery.yml so input_project_dir is set - Accept SU1-SU100 (case-insensitive) in addition to grp0-grp100 for GROUP_NAME - Fix Magellan message to use list format (avoids \n in debug output) - Remove escaped quotes from discovery usage examples Generated with [Devin](https://cli.devin.ai/docs) Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> * fix: extend SU group name support to build_image validation and schemas - Add build_aarch_image tag to input_file_inventory so build_image_aarch64.yml runs provision_config validation (was missing, causing no validation to run for aarch64 builds) - Update GROUP_NAME patterns in functional_groups_config.json and omnia_config.json schemas to accept SU1-SU100 format alongside grp0-grp100 - Update INVALID_GROUP_NAME_MSG to reflect both accepted formats Generated with [Devin](https://cli.devin.ai/docs) Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> --------- Signed-off-by: Sujit Jadhav <sujit_jadhav@dell.com> Co-authored-by: Super User <root@testbed.omnia.test> Co-authored-by: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> * Cleanup discovery roles: move library modules, remove unused roles (#4261) * Cleanup discovery roles: move library modules, remove unused roles - Move ome_server_inventory.py and generate_pxe_mapping.py from discovery/roles/ome_discovery/library/ to common/library/modules/ so they are shared via the common module search path already configured in discovery/ansible.cfg - Remove unused discovery/roles/telemetry/ directory - Remove unused discovery/roles/discovery_validations/ directory - Load discovery_config.yml at playbook level in discovery.yml (consistent with how build_stream_config.yml is loaded in provision.yml) - Fix discovery_complete_msg formatting for readable Ansible output Generated with [Devin](https://cli.devin.ai/docs) Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> * Remove unused discovery_validations role Generated with [Devin](https://cli.devin.ai/docs) Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> --------- Co-authored-by: Super User <root@testbed.omnia.test> Co-authored-by: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> * fix for set_pxe_boot.yml when custom inventory given (#4260) * Update generate_bmc_inventory.yml Signed-off-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> * Update pre_checks.yml Signed-off-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> * lint issue Signed-off-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> --------- Signed-off-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> --------- Signed-off-by: pullan1 <sudha.pullalaravu@dell.com> Signed-off-by: Sujit Jadhav <sujit_jadhav@dell.com> Signed-off-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> Co-authored-by: pullan1 <sudha.pullalaravu@dell.com> Co-authored-by: snarthan <narthan.s@dell.com> Co-authored-by: Sujit Jadhav <sujit_jadhav@dell.com> Co-authored-by: Super User <root@testbed.omnia.test> Co-authored-by: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> Co-authored-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> * resolving merge conflict * revert openchami commit id * resolving review comments * addressing review comments * fix for vmagent scraping powerscale metrics * cleanup script correction for powerscale telemetry cleanup * victoria operator and victoria log input validation * vitoria log input and input validation * remving L2 vslidation for victoria log which is not required * input validation and review comment addressing * change idrac_telemetry_collection_type to telemetry_collection_type * Remove invisible Unicode LRM (U+200E) characters from victoria-operator template filenames * VictoriaLogs container image references and default variable * port check * resolve merge conflict * correction for schema * Update telemetry_config.json * Update validate_input.py * merge conflict telemetry_prereq.yml * change victoria_configurations to victoria_metrics_configurations * remove deployment mode input variable * update for upgrade scenarios * update comments * update comment * resolving issues due to merge conflict * vitoria log changes * victoria log cluster component and VLAgent deployment * updating pod name * removing the changes of adding cert * victoria log changes * remivng victoria log pod calidation playbook * cleanup changes for victoria log * Update ansible-lint.yml and pylint for pub/telemetry (#4296) * Update ansible-lint.yml Signed-off-by: Kratika Patidar <Kratika.Patidar@dell.com> * Update pylint.yml Signed-off-by: Kratika Patidar <Kratika.Patidar@dell.com> * fixing ansible-lint * lint * line-lenght --------- Signed-off-by: Kratika Patidar <Kratika.Patidar@dell.com> --------- Signed-off-by: pullan1 <sudha.pullalaravu@dell.com> Signed-off-by: Sujit Jadhav <sujit_jadhav@dell.com> Signed-off-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> Signed-off-by: priti-parate <140157516+priti-parate@users.noreply.github.com> Signed-off-by: Kratika Patidar <Kratika.Patidar@dell.com> Co-authored-by: mithileshreddy04 <mithilesh.reddy@dell.com> Co-authored-by: priti-parate <140157516+priti-parate@users.noreply.github.com> Co-authored-by: pullan1 <sudha.pullalaravu@dell.com> Co-authored-by: snarthan <narthan.s@dell.com> Co-authored-by: Sujit Jadhav <sujit_jadhav@dell.com> Co-authored-by: Super User <root@testbed.omnia.test> Co-authored-by: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> Co-authored-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> Co-authored-by: Kratika_Patidar <Kratika.Patidar@dell.com> * IB nic ip assignment * update MinIO and registry images to fixed tagged versions, omnia core container tag and version to 2.2 and v2.2.0.0 (#4309) * Minimal OS-only functional group enablement for x86_64 and aarch64 * Update image_package_collector.py * Update provision_validation.py * Minimal OS functional group updates in provision * Minimal OS functional group upgrade * Fix os_* package cross-contamination and remove stale discovery templates * OpenCHAMI upgrade changes * Update openchami container tags * Update main.yml * Update main.yml * Update main.yml * Update omnia version and core tag * vast client installation * single template way * fix(OMN01D-2164): prompt OME credentials only when enable_bmc_discovery is true - Add enable_bmc_discovery flag (default: false) to discovery_config.yml - Load discovery_config.yml in prepare_oim.yml and set ome_discovery_enabled based on enable_bmc_discovery flag - Change discovery credentials from mandatory to conditional_mandatory gated on ome_discovery_enabled - When enable_bmc_discovery is false, OME username/password prompts are skipped during prepare_oim even if ome_ip is pre-filled * fix(OMN01D-2168): fail explicitly when discovery_mechanism is not provided Replace meta: end_play with ansible.builtin.fail so the playbook exits with non-zero status and a clear error message when discovery_mechanism is missing, instead of silently succeeding. * fix(OMN01D-2169): add L1 input validation for discovery.yml - Create discovery_config.json schema with: - enable_bmc_discovery (boolean, required) - ome_ip (string, required; must be valid IPv4 when enable_bmc_discovery is true) - Register discovery_config in config.py files dict and input_file_inventory - Add 'discovery' tag and invoke validate_config.yml in discovery.yml before role execution (consistent with provision.yml pattern) - Add explicit ome_ip check before OME role inclusion for clear fail-fast error when ome_ip is empty with discovery_mechanism=ome * fix(OMN01D-2225): improve OME authentication and reachability error messages - ome_server_inventory.py: auth failure now tells user to verify ome_username/ome_password in omnia_config_credentials.yml and rerun - collect_inventory.yml: wrap wait_for in block/rescue so timeout gives actionable message pointing to discovery_config.yml and network check * fix(OMN01D-2226): correct discovery completion message next steps - Replace misleading 'Rename or copy' instruction with guidance to update pxe_mapping_file_path in provision_config.yml - Show full absolute path of generated file throughout - Add spacing between steps for readability * fix(OMN01D-2227): escape backslash in docstring to suppress SyntaxWarning Python 3.12+ warns about invalid escape sequence '\d' in non-raw string literals. The docstring in extract_su_from_hostname() contained (?=R\d+) which triggered SyntaxWarning during discovery execution. Escaped the backslash to (?=R\\d+) in the docstring. * fix(OMN01D-2230): correct GROUP_NAME and PARENT_SERVICE_TAG in PXE mapping Issue 1 - GROUP_NAME: - Add fallback: try extracting SU pattern from OME group name when BMC hostname has no SU pattern (covers hierarchical OME groups like SU1_slurm_node, SU2_compute, etc.) - grp0 remains the correct default for single-cluster environments Issue 2 - PARENT_SERVICE_TAG: - Define CHILD_ROLES_OF_CONTROL_PLANE set (service_kube_node_x86_64) - Only assign PARENT_SERVICE_TAG to rows whose FUNCTIONAL_GROUP_NAME is a child role of the control plane within the same GROUP_NAME - Control plane nodes, slurm nodes, login nodes, etc. no longer get an incorrect PARENT_SERVICE_TAG * fix(OMN01D-2231): detect and fail on duplicate OME static group assignments - build_device_group_map() now tracks all group memberships per device and returns a conflicts dict for devices in multiple static groups - main() fails with an actionable error listing each conflicting device and its groups, instead of silently using the first-seen group - Prevents incorrect FUNCTIONAL_GROUP_NAME override in PXE mapping * fix(OMN01D-2232): validate OME group names against supported functional groups - Define SUPPORTED_FUNCTIONAL_GROUPS set matching Omnia's known roles - Skip servers whose OME static group is not in the supported set - Emit a warning per skipped device listing the unsupported group name and the full set of supported groups - Unsupported groups (e.g. 'abc') no longer appear in the PXE mapping * revert: temporarily revert discovery fixes for PR workflow * s3cmd configurations * update image tags * fix: Resolve FK constraint violation and catalog metadata persistence issues - Fix FK constraint by using shared session for image_group_repo and image_repo in ResultPoller - Concatenate multiple S3 paths with semicolon delimiter to respect unique constraint uq_images_image_group_id_role - Update cleanup_job.py to split semicolon-delimited paths and delete each individually - Widen image_name column from VARCHAR(256) to VARCHAR(512) to accommodate semicolon-delimited paths - Add logging to FileArtifactStore.store() to debug file write failures - Fix URI construction in parse_catalog.py for existing artifacts (construct file:// URI directly) - Update build_stream container image tag to 1.1 Tested with job 7da54d1f-ed26-41dd-b3ff-0386104e644d (image-build19) - ImageGroup and Images created successfully * Fix lint error * Modify the charlimit to 512 in DB * Fix the image_groups name and stages in summary * Updating Catalog changes aligning to recent updates * Fix the pylint issues * Catalog Updates to support powerscale drivers --------- Signed-off-by: balajikumaran.cs <balajikumaran.c.s@gmail.com> Signed-off-by: pullan1 <sudha.pullalaravu@dell.com> Signed-off-by: Sujit Jadhav <sujit_jadhav@dell.com> Signed-off-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> Signed-off-by: priti-parate <140157516+priti-parate@users.noreply.github.com> Signed-off-by: Kratika Patidar <Kratika.Patidar@dell.com> Co-authored-by: Mithilesh Reddy <mithilesh.reddy@dell.com> Co-authored-by: mcas <sakshi.s@dell.com> Co-authored-by: Super User <root@oim.omnia.test> Co-authored-by: snarthan <narthan.s@dell.com> Co-authored-by: balajikumaran.cs <balajikumaran.c.s@gmail.com> Co-authored-by: balajikumaran-c-s <balajikumaran.cs@dellteam.com> Co-authored-by: priti-parate <140157516+priti-parate@users.noreply.github.com> Co-authored-by: Abhishek S A <abhishek.sa3@dell.com> Co-authored-by: pullan1 <sudha.pullalaravu@dell.com> Co-authored-by: Sujit Jadhav <sujit_jadhav@dell.com> Co-authored-by: Super User <root@testbed.omnia.test> Co-authored-by: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> Co-authored-by: SOWJANYAJAGADISH123 <Sowjanya.Jagadish@dell.com> Co-authored-by: Kratika_Patidar <Kratika.Patidar@dell.com> Co-authored-by: Nagachandan-P <Nagachandan.p@dell.com> Co-authored-by: Sujit Jadhav <sujit.jadhav@dell.com>

* vitoria log input and input validation * remving L2 vslidation for victoria log which is not required * input validation and review comment addressing * change idrac_telemetry_collection_type to telemetry_collection_type * Remove invisible Unicode LRM (U+200E) characters from victoria-operator template filenames * VictoriaLogs container image references and default variable * port check * resolve merge conflict * correction for schema * Update telemetry_config.json * Update validate_input.py * merge conflict telemetry_prereq.yml * change victoria_configurations to victoria_metrics_configurations * remove deployment mode input variable * update for upgrade scenarios * update comments * update comment * resolving issues due to merge conflict * vitoria log changes * adding the cuda and nvidia driver version gate check * victoria log cluster component and VLAgent deployment * updating pod name * storage_config.yml added * Added logic for mounts * removing the changes of adding cert * victoria log changes * remivng victoria log pod calidation playbook * cleanup changes for victoria log * Update ansible-lint.yml Signed-off-by: Kratika Patidar <Kratika.Patidar@dell.com> * Update pylint.yml Signed-off-by: Kratika Patidar <Kratika.Patidar@dell.com> * fixing ansible-lint * lint * line-lenght * Update ansible-lint.yml and pylint for pub/telemetry (#4296) * Update ansible-lint.yml Signed-off-by: Kratika Patidar <Kratika.Patidar@dell.com> * Update pylint.yml Signed-off-by: Kratika Patidar <Kratika.Patidar@dell.com> * fixing ansible-lint * lint * line-lenght --------- Signed-off-by: Kratika Patidar <Kratika.Patidar@dell.com> * adding SAN for vicotir agent * powerscale authorization input and input validation for powrscale authorization and powerscale telemetry (#4288) * input validation for powrscale authorization and powerscale telemetry * remove unused messages * fix cide review comments * fix UT issues * input valdiation onlu when powerscale authorization is true * pylint fixes * pylint fixes * update copyright year * Updating SAN changes for vmagengt and vlagent * Updating the SAN for vlagent (#4298) * telemetry input restructure Signed-off-by: Abhishek S A <abhishek.sa3@dell.com> * update telemetry * Update victorialogs-operator-vlcluster.yaml.j2 * Update telemetry.sh.j2 * telemetry input restructure (#4299) * telemetry input restructure Signed-off-by: Abhishek S A <abhishek.sa3@dell.com> * update telemetry * Update victorialogs-operator-vlcluster.yaml.j2 * Update telemetry.sh.j2 --------- Signed-off-by: Abhishek S A <abhishek.sa3@dell.com> * build image fix * build image suffix * Update service_k8s.json Signed-off-by: Kratika Patidar <Kratika.Patidar@dell.com> * issue fix * issue fix * build image fix * Revert "issue fix" This reverts commit 847746c. * Revert changes in telemetry branch (#4308) * Revert "Merge branch 'pub/telemetry' into pub/telemetry" This reverts commit 9b8faf8, reversing changes made to 895354b. * update telemetry fix * Mounts level 1 complete * initial cuda dnf installation * updating arch files * changes with respect to new input structure * powerscale_metrics_enabled variable access * powerscale telemetry enabled metrics access * pylint fixes * changes with respect to new input structure (#4312) * changes with respect to new input structure * powerscale_metrics_enabled variable access * powerscale telemetry enabled metrics access * pylint fixes * vlagent with loadbalancer type * feat(mount_config): enhance per-node bind mounts and secure runcmd execution - Implement literal moustache preservation for cloud-init variables using !unsafe with from_yaml. - Generalize node_key support to allow full metadata paths (e.g., ds.meta_data.instance_data.local_ipv4). - Fix b64encode evaluation in permission and directory creation tasks to ensure valid shell command generation. - Centralize mount-specific runcmd execution in cloud-init templates using base64 encoding to prevent shell injection/parsing issues. - Update storage_config schema to support flexible cloud-init variable references for node_key. - Add safety defaults for mount_on_oim selection to prevent 'attribute not found' errors. - Restore 'union' filter for mounts to maintain idempotency across multiple runs. * remove 9429 as it get added by operator * updating input validation, and default vars for vecotr * remove trailing space * feat(storage,provision): add cloud-init swap support + schema/validation updates New tasks: provision/roles/mount_config/tasks/{swap_config.yml,process_single_swap.yml} Build per-functional-group swap entries in cloud_init_groups_dict Validate swap filename path and size format; include optional maxsize Idempotent merge via combine(recursive=True) Render swap block in cloud-init group templates Inject conditional swap: {filename, size, [maxsize]} across login, compiler, kube (control/node), slurm (control/node) for x86_64/aarch64 Only rendered when cloud_init_groups_dict[fg].swap is present Strengthen input validation (common_validation.py) Enforce no overlap of functional_group_prefix/group across swap entries Validate maxsize >= size (supports bytes or K/M/G/T, and “auto”) Add helper parsers and validators for size comparison Update storage_config schema for swap (storage_config.json) Remove deprecated swap.name field Required: filename, size oneOf: require either functional_group_prefix or group Preserve existing size semantics (supports units/auto) * powerscale syslog support * update copyright year * UT issues * remove TLS mode as it is not supported forsyslog on VLAgent * make syslog source IP as optional * UT issue fixes * VLagent to victorial log transformation * VLAgent to victorial logs cert issue * updating default version * remving vector-idrac * fix for the script does not push the new cert to the Kubernetes secret or restart pods * vast client installation * refactor: make determine_target_groups.yml generic for mounts and swap - Change mount_item reference to config_item parameter - Update process_single_mount.yml to pass config_item: mount_item - Update process_single_swap.yml to pass config_item: swap_item - Enable reuse of target group determination logic This allows both mount and swap processing to use the same task. * single template way * Certificate resolution issue * rsyslog configuration * Added powervault to all * loadbalancer configuration * Added powervault to all functional groups * logs with loadbalncer configuration * get syslog with nodeport * syslog collection work on nodeport * ansible lint fix * update remotewrite url * Verify script * removing audit trail log * login compiler node cuda lock installation * updated the message for nodes where installation skips * Powervaulr iscsi packages added to all nodes * fix(OMN01D-2164): prompt OME credentials only when enable_bmc_discovery is true - Add enable_bmc_discovery flag (default: false) to discovery_config.yml - Load discovery_config.yml in prepare_oim.yml and set ome_discovery_enabled based on enable_bmc_discovery flag - Change discovery credentials from mandatory to conditional_mandatory gated on ome_discovery_enabled - When enable_bmc_discovery is false, OME username/password prompts are skipped during prepare_oim even if ome_ip is pre-filled * fix(OMN01D-2168): fail explicitly when discovery_mechanism is not provided Replace meta: end_play with ansible.builtin.fail so the playbook exits with non-zero status and a clear error message when discovery_mechanism is missing, instead of silently succeeding. * fix(OMN01D-2169): add L1 input validation for discovery.yml - Create discovery_config.json schema with: - enable_bmc_discovery (boolean, required) - ome_ip (string, required; must be valid IPv4 when enable_bmc_discovery is true) - Register discovery_config in config.py files dict and input_file_inventory - Add 'discovery' tag and invoke validate_config.yml in discovery.yml before role execution (consistent with provision.yml pattern) - Add explicit ome_ip check before OME role inclusion for clear fail-fast error when ome_ip is empty with discovery_mechanism=ome * fix(OMN01D-2225): improve OME authentication and reachability error messages - ome_server_inventory.py: auth failure now tells user to verify ome_username/ome_password in omnia_config_credentials.yml and rerun - collect_inventory.yml: wrap wait_for in block/rescue so timeout gives actionable message pointing to discovery_config.yml and network check * fix(OMN01D-2226): correct discovery completion message next steps - Replace misleading 'Rename or copy' instruction with guidance to update pxe_mapping_file_path in provision_config.yml - Show full absolute path of generated file throughout - Add spacing between steps for readability * fix(OMN01D-2227): escape backslash in docstring to suppress SyntaxWarning Python 3.12+ warns about invalid escape sequence '\d' in non-raw string literals. The docstring in extract_su_from_hostname() contained (?=R\d+) which triggered SyntaxWarning during discovery execution. Escaped the backslash to (?=R\\d+) in the docstring. * fix(OMN01D-2230): correct GROUP_NAME and PARENT_SERVICE_TAG in PXE mapping Issue 1 - GROUP_NAME: - Add fallback: try extracting SU pattern from OME group name when BMC hostname has no SU pattern (covers hierarchical OME groups like SU1_slurm_node, SU2_compute, etc.) - grp0 remains the correct default for single-cluster environments Issue 2 - PARENT_SERVICE_TAG: - Define CHILD_ROLES_OF_CONTROL_PLANE set (service_kube_node_x86_64) - Only assign PARENT_SERVICE_TAG to rows whose FUNCTIONAL_GROUP_NAME is a child role of the control plane within the same GROUP_NAME - Control plane nodes, slurm nodes, login nodes, etc. no longer get an incorrect PARENT_SERVICE_TAG * fix(OMN01D-2231): detect and fail on duplicate OME static group assignments - build_device_group_map() now tracks all group memberships per device and returns a conflicts dict for devices in multiple static groups - main() fails with an actionable error listing each conflicting device and its groups, instead of silently using the first-seen group - Prevents incorrect FUNCTIONAL_GROUP_NAME override in PXE mapping * fix(OMN01D-2232): validate OME group names against supported functional groups - Define SUPPORTED_FUNCTIONAL_GROUPS set matching Omnia's known roles - Skip servers whose OME static group is not in the supported set - Emit a warning per skipped device listing the unsupported group name and the full set of supported groups - Unsupported groups (e.g. 'abc') no longer appear in the PXE mapping * fix(OMN01D-2168): fail explicitly when discovery_mechanism is not provided Replace meta: end_play with ansible.builtin.fail so the playbook exits with non-zero status and a clear error message when discovery_mechanism is missing, instead of silently succeeding. * fix(OMN01D-2169): add L1 input validation for discovery.yml - Create discovery_config.json schema with: - enable_bmc_discovery (boolean, required) - ome_ip (string, required; must be valid IPv4 when enable_bmc_discovery is true) - Register discovery_config in config.py files dict and input_file_inventory - Add 'discovery' tag and invoke validate_config.yml in discovery.yml before role execution (consistent with provision.yml pattern) - Add explicit ome_ip check before OME role inclusion for clear fail-fast error when ome_ip is empty with discovery_mechanism=ome * fix(OMN01D-2225): improve OME authentication and reachability error messages - ome_server_inventory.py: auth failure now tells user to verify ome_username/ome_password in omnia_config_credentials.yml and rerun - collect_inventory.yml: wrap wait_for in block/rescue so timeout gives actionable message pointing to discovery_config.yml and network check * fix(OMN01D-2226): correct discovery completion message next steps - Replace misleading 'Rename or copy' instruction with guidance to update pxe_mapping_file_path in provision_config.yml - Show full absolute path of generated file throughout - Add spacing between steps for readability * fix(OMN01D-2227): escape backslash in docstring to suppress SyntaxWarning Python 3.12+ warns about invalid escape sequence '\d' in non-raw string literals. The docstring in extract_su_from_hostname() contained (?=R\d+) which triggered SyntaxWarning during discovery execution. Escaped the backslash to (?=R\\d+) in the docstring. * fix(OMN01D-2230): correct GROUP_NAME and PARENT_SERVICE_TAG in PXE mapping Issue 1 - GROUP_NAME: - Add fallback: try extracting SU pattern from OME group name when BMC hostname has no SU pattern (covers hierarchical OME groups like SU1_slurm_node, SU2_compute, etc.) - grp0 remains the correct default for single-cluster environments Issue 2 - PARENT_SERVICE_TAG: - Define CHILD_ROLES_OF_CONTROL_PLANE set (service_kube_node_x86_64) - Only assign PARENT_SERVICE_TAG to rows whose FUNCTIONAL_GROUP_NAME is a child role of the control plane within the same GROUP_NAME - Control plane nodes, slurm nodes, login nodes, etc. no longer get an incorrect PARENT_SERVICE_TAG * fix(OMN01D-2231): detect and fail on duplicate OME static group assignments - build_device_group_map() now tracks all group memberships per device and returns a conflicts dict for devices in multiple static groups - main() fails with an actionable error listing each conflicting device and its groups, instead of silently using the first-seen group - Prevents incorrect FUNCTIONAL_GROUP_NAME override in PXE mapping * fix(OMN01D-2232): validate OME group names against supported functional groups - Define SUPPORTED_FUNCTIONAL_GROUPS set matching Omnia's known roles - Skip servers whose OME static group is not in the supported set - Emit a warning per skipped device listing the unsupported group name and the full set of supported groups - Unsupported groups (e.g. 'abc') no longer appear in the PXE mapping * revert: temporarily revert discovery fixes for PR workflow * fix(OMN01D-2164): prompt OME credentials only when enable_bmc_discovery is true - Add enable_bmc_discovery flag (default: false) to discovery_config.yml - Load discovery_config.yml in prepare_oim.yml and set ome_discovery_enabled based on enable_bmc_discovery flag - Change discovery credentials from mandatory to conditional_mandatory gated on ome_discovery_enabled - When enable_bmc_discovery is false, OME username/password prompts are skipped during prepare_oim even if ome_ip is pre-filled * slurm cuda,dcg,peermem lock installation * vast enabled flag * fix: remove unsupported service_kube_control_plane_first_x86_64 from SUPPORTED_FUNCTIONAL_GROUPS * adding mod[rpbe for peermem * Victoria defect fixes * victoria changes * fix: Add duplicate mount point validation per functional group - Fix pxe_mapping_file reference in config and validation flow - Add validation to detect duplicate mount points per expanded functional group - Include entry names in error messages for clarity - Update schema and examples for consistency * Revert "revert the vast enabled flag" This reverts commit d20d1fb. * vast conditional installation * refactor: Align storage_config.yml with authoritative schema - Remove extra mount_params profiles (vast_nfs_performance, powervault_iscsi, network_storage, local_storage, bind_mounts, scratch_storage, global) - Keep only essential profiles: default and vast_nfs - Simplify mounts section with two core examples (nfs_slurm, nfs_k8s) - Remove VAST-specific entries and per-node bind mount examples - Comment out powervault_config entries (move to examples) - Comment out swap section (move to examples) - Fix field names: group → groups in comments and examples - Remove vast_enabled flag - Align with storage_config.json schema (only mount_params, mounts, powervault_config, swap sections) - Update comments to reference correct field names (groups instead of group) * comment updated configure-ib-network.sh.j2 Signed-off-by: Nagachandan P <Nagachandan.p@dell.com> * Revereted kube node changes * Removed acidentally committed files * Added ansible.cfg * Ansible lint fixes * lint issue Signed-off-by: Nagachandan P <Nagachandan.p@dell.com> * dns for ib network * adding cuda mount for slurm * adding the dcgm_support * adding variable approach * merge conflict resolution * revert condition * input comment * Adding new files * ansible lint error fixes * feat(discovery): robust OME pagination, GPU collection, and timestamped PXE mapping - Replace @odata.nextLink-based pagination with explicit $top/$skip OData parameters for deterministic, scalable device retrieval (8k-20k+ nodes) - Add configurable page_size parameter (default 200, max 1000) to ome_server_inventory module and wire through Ansible role defaults - Add retry with exponential backoff for transient HTTP 5xx/timeout failures - Return pagination stats (total_devices, page_size, total_pages, pages_fetched) in module output; display summary in playbook - Collect GPU vendor and type from OME devicePciDevice inventory; append GPU_VENDOR and GPU_TYPE columns to PXE mapping CSV - Generate timestamped PXE mapping files (bmc_pxe_mapping_file_<ts>.csv) so each run produces a unique file - Handle OME returning list instead of dict from inventory endpoints - Fix telemetry_config.yml validation: point to telemetry_validation module instead of removed common_validation.validate_telemetry_config - Add 20 unit tests covering pagination, retry, clamping, and filtering * variable chnages wrt new telemetry_config.yml chnages * fix(lint): resolve yaml line-length violations (>160 chars) * feat: implement host-specific mount map in cloud-init templates - Enable host-specific mount map generation from PXE mapping data - Add vendor_data metadata structure to cloud-init runcmd for dynamic mounts - Update cloud-init variable syntax from Jinja2 template to cloud-init query - Fix variable reference in build_host_mount_map.yml (PXE_GROUP -> GROUP_NAME) - Change mkdir commands to verbose mode (mkdir -pv) - Update pxe_mapping_file_path to use omnia_shared_path with path normalization This allows per-host mount configurations to be dynamically applied during node provisioning via cloud-init vendor_data metadata. * Removed group map removed legacy code * Update logical_validation.py- remove repetitive import * Update telemetry_pod_cleanup.yaml.j2 - revert telemetry pod cleanup changes * Update idrac_telemetry_statefulset.yaml.j2 - revert statefulset changes * Update telemetry_prereq.yml-remove unnecessary comment * Update csi_driver_powerscale.json- update URL * Update service_k8s.json-update version as per release 1.16.3 * Update telemetry_config.yml- fix csm observability values.yaml version * Update idrac_telemetry_statefulset.yaml.j2-update variable parsing * Update configure_powerscale_syslog.sh.j2 -fix syntax error for syslog * removing dcgm_support variable * removing stale entries * remove default dns entry for ib network Signed-off-by: Nagachandan P <Nagachandan.p@dell.com> * image layer initial commit * minor chnages * mounting hpc_tools * removinf from control node * Fix mount ordering and bind mount creation for cloud-init runcmd - Restructure runcmd to mount NFS shares before creating bind mount sources - Add explicit mount commands after each NFS fstab entry to ensure mounts are available before subdirectories are created on them - Add mkdir commands for bind mount target directories on mounted NFS - Consolidate mount -av at appropriate checkpoints to avoid premature mounting before all fstab entries and directories are ready - Update process_single_mount.yml to add mkdir for bind mount targets before fstab entries - Add "Refresh mounts" task to append final mount -av after all setup This ensures the correct sequence: 1. Add NFS mount to fstab 2. Mount NFS share 3. Create subdirectories on mounted NFS 4. Add bind mount entries to fstab 5. Mount bind mounts Fixes mount failures: "special device does not exist" and "mount point does not exist" errors during cloud-init execution. * VMscraper to scrape metrics from OTEL collector (#4362) * pub telemetry changes * service to scrape metrics from OTEL collector * vmservice to scrape metrics from otel collector * update endpoints * revert other changes * revert merge changes as per head * revert variable set * revert changes * revert changes * pylint fixes * ansible lint fixes * updating completion messaage * telemetry validation while prepare oim * update condition * added check for LDMS * Fix: Add retry logic for ochami node discovery with SMD readiness checks - Extract service readiness checks into separate block/rescue pattern - Add SMD API health check before discovery attempt - Implement automatic retry on discovery failure with service restart - Increase service check timeout from 2 to 2 minutes (12 retries × 10s delay) - Prevent connection refused errors by ensuring SMD endpoint is ready This addresses the race condition where systemd marks smd service as "started" but the HTTP endpoint at oimcp.oim.test:8443 isn't accepting connections yet. * Update smd api check command * chnaged approach * Cloud_init and other service failure messages * Updated messages * fix: BMC discovery credential prompting and PARENT_SERVICE_TAG assignment (#4366) 1. Skip OME credential prompt when enable_bmc_discovery is false - Changed discovery credentials from mandatory to conditional_mandatory in credential utility vars (gated on enable_bmc_discovery) - Added set_fact in prepare_oim.yml to promote enable_bmc_discovery from namespaced to top-level scope before credential utility runs 2. Fix PARENT_SERVICE_TAG assignment in PXE mapping - Source changed from service_kube_control_plane to service_kube_node - Only slurm_node_aarch64 and slurm_node_x86_64 receive PARENT_SERVICE_TAG - All other roles (control_plane, kube_node, login, slurm_control) remain empty * changing the script a little * ansible lint * ansible lint * Update provision_mapping_nodes.yml Signed-off-by: Jagadeesh N V <39791839+jagadeeshnv@users.noreply.github.com> * Refactor: Update K8s NFS configuration to use unified mounts structure - Replace nfs_client_params with mounts list for consistency - Extract k8s_nfs_server_ip from mounts source field (split on ':') - Extract k8s_client_mount_path from mounts mount_point field - Introduce k8s_nfs_server_path as unified source reference - Update error message to use k8s_nfs_server_path for clarity - Comment out deprecated k8s_server_share_path and nfs_server_ip references - Consolidate 'mount -av' into single set_fact call instead of separate task - Remove redundant debug task for single_mnt_runcmd - Add loop_control label for better Ansible output readability - Improve code formatting for regex_replace operations (no functional change) - Streamline permission and bind mount runcmd list construction * Updated xlurm custom repo names * Updated repo names * Removed static mounts block * rearrange args * Removed unecessary mounts * Merge pull request #4371 from Kratika-P/pub/q2_dev Telemetry enabled flag set for creating nd executing the telemetry.sh * fix(provision): resolve K8s NFS mount race condition and improve Slurm support detection - Move mkdir for K8s NFS mount directories before fstab entries to prevent race condition where mount attempts occur before target directories exist - Fix duplicate mount point names in storage_config.yml (ps3: slurm_login -> slurm_login_kube_node, ps4: login_node -> login_kubectrl) to avoid conflicts between functional groups - Add mkdir -p for /tmp/crio-storage, /var/lib/etcd, /var/lib/kubelet, /etc/kubernetes, /var/log/pods, /var/lib/packages BEFORE adding NFS entries to /etc/fstab - Affected templates: - ci-group-service_kube_control_plane_first_x86_64.yaml.j2 - ci-group-service_kube_control_plane_x86_64.yaml.j2 - ci-group-service_kube_node_x86_64.yaml.j2 - Remove redundant 'mount -av' from bind mount runcmd sequence (process_single_mount.yml) - Functional groups now handle mount -av independently - Fix slurm_support detection logic to check both software_config AND functional_groups - Initialize slurm_support to false, then set true only if: 1. slurm_custom is in software_config.softwares, AND 2. slurm_control_node_x86_64 exists in functional_groups - Apply fix to both mount_config and slurm_config roles - Remove unused slurm_support var from mount_config/vars/main.yml - Remove verbose flags from swap commands (fallocate -v, chmod -v, mkswap -v, swapon -v) in ci-group-login_compiler_node_aarch64.yaml.j2 - Update storage_config.json schema to reflect new mount point naming conventions This commit addresses Issue #3 (race condition) and Issue #4 (missing directory creation) from the cloud-init analysis, preventing silent NFS mount failures on K8s nodes. * Added missing mount-a * papi url chnaged * Updated powervault to service_kube nodes * LDAP home mount updated * Input refinement * Updated descriptions * gpu defect- peermem defect- prolog removal * defect fix for pxe validation * lint fixes * Templated swap part --------- Signed-off-by: priti-parate <140157516+priti-parate@users.noreply.github.com> Signed-off-by: Kratika Patidar <Kratika.Patidar@dell.com> Signed-off-by: Abhishek S A <abhishek.sa3@dell.com> Signed-off-by: sakshi-singla-1735 <sakshi.s@dell.com> Signed-off-by: Jagadeesh N V <39791839+jagadeeshnv@users.noreply.github.com> Signed-off-by: Sujit Jadhav <sujit_jadhav@dell.com> Signed-off-by: Nagachandan P <Nagachandan.p@dell.com> Co-authored-by: priti-parate <140157516+priti-parate@users.noreply.github.com> Co-authored-by: Kratika_Patidar <Kratika.Patidar@dell.com> Co-authored-by: mcas <sakshi.s@dell.com> Co-authored-by: Jagadeesh N V <jagadeesh_n_v@dell.com> Co-authored-by: Nagachandan-P <Nagachandan.p@dell.com> Co-authored-by: snarthan <narthan.s@dell.com> Co-authored-by: Sujit Jadhav <sujit.jadhav@dell.com> Co-authored-by: Jagadeesh N V <39791839+jagadeeshnv@users.noreply.github.com> Co-authored-by: Sujit Jadhav <sujit_jadhav@dell.com>

Kratika-P added 2 commits April 21, 2026 10:42

Update ansible-lint.yml

99d466f

Signed-off-by: Kratika Patidar <Kratika.Patidar@dell.com>

Update pylint.yml

a61950e

Signed-off-by: Kratika Patidar <Kratika.Patidar@dell.com>

Kratika-P changed the title ~~Update ansible-lint.yml~~ Update ansible-lint.yml and pylint for pub/telemetry Apr 21, 2026

abhishek-sa1 approved these changes Apr 21, 2026

View reviewed changes

priti-parate approved these changes Apr 21, 2026

View reviewed changes

Kratika-P added 3 commits April 21, 2026 05:42

fixing ansible-lint

3905660

lint

101efee

line-lenght

eb77a81

abhishek-sa1 merged commit d8e2ffe into dell:pub/telemetry Apr 21, 2026
3 checks passed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Update ansible-lint.yml and pylint for pub/telemetry#4296

Update ansible-lint.yml and pylint for pub/telemetry#4296
abhishek-sa1 merged 5 commits into
dell:pub/telemetryfrom
Kratika-P:pub/telemetry

Kratika-P commented Apr 21, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Conversation

Kratika-P commented Apr 21, 2026

Description of the Solution

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants