-
Notifications
You must be signed in to change notification settings - Fork 104
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Existing Fmwcluster is not detected therefore it proceeds to create OPSS store #49
Comments
Edwin, thanks in advance for looking into this. The aim is to minimize the config. changes if possible in the upgrade as we have quite a few SOA domains and to minimize regression testing. |
Hi, Strange, Ok there can be 2 issues facter can fail because it can't read the domain and generate facts. Or the soa cluster check fails Can you run in debug mode and send me the first output , this analyze the current By the way when you use a puppet master you should restart the puppet master services because it can cache some old puppet code. Thanks |
And you should have this [vagrant@soa2admin ~]$ sudo cat /etc/wls_domains.yaml
and this when you start a new provision step
|
Thanks Edwin, I've had to back out the change as it was a weekend upgrade and we ran out of time. I will post the debug when I get a chance for "Upgrade Round 2" in a few weeks time. Many thanks... |
Hi Edwin... Can you please re-open this issue? Re-trying the upgrade to the latest: biemond-orawls-1.0.17 Double checked /etc/wls_domains.yaml and looks fine: domains: Looking at the facts, I noticed that before the upgrade: ora_mdw_domain_0_soa => soa_app_cluster After the upgrade I get: ora_mdw_domain_0_soa => NotFound Much appreciated if you can help with this... Thanks, Jason |
interesting, maybe you can add some extra debug to https://github.com/biemond/biemond-orawls/blob/master/lib/facter/orawls.rb I think somewhere in reading the domain config.xml file goes totally wrong. |
Hi Edwin, NOTE, that the cluster and node managers are running when I update the Puppet modules. Should I stop all java processes before adding the new Puppet modules, then do a Puppet run... Do you have any suggested best practice for upgrading the Puppet modules? Thanks... |
No need to shutdown, I will detect this and won't do an upgrade or change something unless you make a change to the model. more important after upgrade of the modules.
|
A search of the Puppet run debug logs for "ora_mdw_domain" returns the below and fails since it thinks it needs to create the OPSS store since it already exists... Notice: /Stage[main]/Profile::Wls::Fmw_cluster/Orawls::Utils::Fmwcluster[appSoaCluster]/Exec[execwlst create OPSS store ipt_app_domain appSoaCluster] [0;36mDebug: orawls.rb ora_mdw_domain_0 /u01/app/oracle/admin/domains/soa_app_domain[0m |
Ok Let's comment out the jrf_cluster entry in the puppet configuration, else we get some corruption can you add some extra debug statements to the orawls.rb facter file I check for soa-infra deployment deployments = ''
end unless soaTargets.nil? |
Thanks...yes, I'll add debug and re-run as per your suggestion...Cheers |
Hi Edwin, I've added debug...can you kindly have a look please? Facts BEFORE first puppet run following Puppet modules upgrade ora_inst_loc_data => /u01/app/oracle/product/oraInventory Puppet Debug from 1st Puppet Run �[0;36mDebug: orawls.rb weblogic user is oracle�[0m �[0;36mDebug: orawls.rb weblogic user is oracle�[0m �[mNotice: /Stage[main]/Profile::Wls::Clusters/Wls_cluster[proj_app_cluster]/servers: servers changed ['app_server1', 'app_server2'] to 'app_server1,app_server2'�[0m �[0;36mDebug: /Stage[main]/Profile::Wls::Clusters/Wls_cluster[proj_app_cluster]: The container Class[Profile::Wls::Clusters] will propagate my refresh event�[0m The below changes appear to break the SOA cluster - soa-infra goes to installed and is NO LONGER targetted to the SOA cluster �[mNotice: /Stage[main]/Profile::Wls::Groups/Wls_group[Administrators]/users: users changed ['weblogic', 'app_deployer', 'sst_tester'] to 'weblogic,app_deployer,sst_tester'�[0m �[mNotice: /Stage[main]/Profile::Wls::Groups/Wls_group[proj_jms_users]/users: users changed ['proj_jms_dev', 'proj_jms_test'] to 'proj_jms_dev,proj_jms_test'�[0m �[mNotice: /Stage[main]/Profile::Wls::Jms_module/Wls_jms_module[PROJJMSSystemResources]/target: defined 'target' as 'proj_app_cluster'�[0m �[mNotice: /Stage[main]/Profile::Wls::Jms_subdeployment/Wls_jms_subdeployment[PROJJMSSystemResources:PROJJMSSubDeployment]/target: target changed ['PROJJMSServer1', 'PROJJMSServer2'] to 'PROJJMSServer1,PROJJMSServer2'�[0m �[mNotice: /Stage[main]/Profile::Wls::Clusters/Wls_cluster[proj_app_cluster]/servers: servers changed ['app_server1', 'app_server2'] to 'app_server1,app_server2'�[0m Facts AFTER first puppet run following Puppet modules upgrade. ora_mdw_domain_0_soa => NotFound --> Each subsequent Puppet run tries to create FMWcluster & OPSS database ora_inst_loc_data => /u01/app/oracle/product/oraInventory Puppet Debug from Subsequent Puppet Runs ALL Fail [0;36mDebug: Custom orawls.rb domain_path /u01/app/oracle/admin/domains/proj_app_domain[0m Notice: /Stage[main]/Profile::Wls::Fmw_cluster/Orawls::Utils::Fmwcluster[appSoaCluster]/File[/home/oracle/installers/soa-bpm-createUDDappSoaCluster.py]/content: content changed '{md5}1459b21c98690a5be5dbbbd34c09296b' to '{md5}bb10efb6b26e5cdd041744d6ffef421f' Notice: /Stage[main]/Profile::Wls::Fmw_cluster/Orawls::Utils::Fmwcluster[appSoaCluster]/Orawls::Control[ShutdownAdminServerForSoaappSoaCluster]/Wls_adminserver[ShutdownAdminServerForSoaappSoaCluster:AdminServer]: Dependency Exec[execwlst create OPSS store proj_app_domain appSoaCluster] has failures: true |
Sorry Edwin for the long comment as I couldn't attach as a log. In summary, it looks like the following changes in the first run breaks the existing SOA cluster configuration. Please NOTE that we're not using your module to install the base weblogic 10.3.6 and soasuite 11.1.1.7 binaries (custom RPMs) and this has worked fine so far until now. The below changes appear to break the SOA cluster - soa-infra goes to installed and is NO LONGER targetted to the SOA cluster [mNotice: /Stage[main]/Profile::Wls::Groups/Wls_group[Administrators]/users: users changed ['weblogic', 'app_deployer', 'sst_tester'] to 'weblogic,app_deployer,sst_tester'[0m [mNotice: /Stage[main]/Profile::Wls::Groups/Wls_group[proj_jms_users]/users: users changed ['proj_jms_dev', 'proj_jms_test'] to 'proj_jms_dev,proj_jms_test'[0m [mNotice: /Stage[main]/Profile::Wls::Jms_module/Wls_jms_module[PROJJMSSystemResources]/target: defined 'target' as 'proj_app_cluster'[0m [mNotice: /Stage[main]/Profile::Wls::Jms_subdeployment/Wls_jms_subdeployment[PROJJMSSystemResources:PROJJMSSubDeployment]/target: target changed ['PROJJMSServer1', 'PROJJMSServer2'] to 'PROJJMSServer1,PROJJMSServer2'[0m [mNotice: /Stage[main]/Profile::Wls::Clusters/Wls_cluster[proj_app_cluster]/servers: servers changed ['app_server1', 'app_server2'] to 'app_server1,app_server2'[0m |
Hi, this will helps. [0;36mDebug: orawls.rb ora_mdw_domain_0_soa proj_app_cluster[0m [mNotice: /Stage[main]/Profile::Wls::Clusters/Wls_cluster[proj_app_cluster]/servers: servers changed ['app_server1', 'app_server2'] to 'app_server1,app_server2'[0m this is strange maybe this breaks it , I changed this to an array instead of a string example of wls_cluster
in hiera
/Stage[main]/Profile::Wls::Fmw_cluster/Orawls::Utils::Fmwcluster[appSoaCluster]/Notify[Custom - soa_cluster_name: proj_app_cluster]: The container Orawls::Utils::Fmwcluster[appSoaCluster] will propagate my refresh event[0m The below changes appear to break the SOA cluster - soa-infra goes to installed and is NO LONGER targetted to the SOA cluster I think this is the problem
I shouldn't do remove removeReferencesToBean(cl), this remove it Can you test this for me by removing this line or get the latest version of github Thanks |
Hi Edwin...its worked. Once again, thank you very much... |
I have to thank you, this was a serious bug and can destroy your domain. |
Hi Edwin
I have an co-located pair of SOA domains/cluster running on two VMs (Centos 6.5). Following upgrade (see below), the existing FmwCluster instances are not detected therefore it proceeds to create OPSS store and therefore breaks the Puppet run and its dependencies.
orawls 1.0.1 --> 1.0.11
orautils 0.2.4 --> 0.2.9
easy_type 0.7.4 --> 0.12.4
Notice: orawls::utils::fmwcluster Soa2Cluster 1036 continue true cause not exists
Notice: /Stage[main]/Profile::Wls::Fmw_cluster/Orawls::Utils::Fmwcluster[Soa2Cluster]/Notify[orawls::utils::fmwcluster Soa2Cluster 1036 continue true cause not exists]/message: defined 'message' as 'orawls::utils::fmwcluster Soa2Cluster 1036 continue true cause not exists'
Notice: /Stage[main]/Profile::Wls::Fmw_cluster/Orawls::Utils::Fmwcluster[Soa2Cluster]/Exec[execwlst create OPSS store soa2_domain Soa2Cluster]/returns: Command FAILED, Reason: oracle.security.jps.service.policystore.PolicyObjectAlreadyExistsException: javax.persistence.PersistenceException: Exception [EclipseLink-4002](Eclipse Persistence Services - 2.3.1.v20111018-r10243): org.eclipse.persistence.exceptions.DatabaseException
Existing YAML:
fmw_cluster_instances:
Soa1Cluster:
domain_name: soa1_domain
adminserver_name: AdminServer
adminserver_address: soa01
adminserver_port: 7001
soa_cluster_name: soa1_cluster
bam_cluster_name: soa1_bam_cluster
log_output: true
bpm_enabled: true
bam_enabled: true
soa_enabled: true
osb_enabled: false
repository_prefix: APP
tag: mwconfig
Soa2Cluster:
domain_name: soa2_domain
adminserver_name: AdminServer
adminserver_address: soa01
adminserver_port: 7101
soa_cluster_name: soa2_cluster
bam_cluster_name: soa2_bam_cluster
log_output: true
bpm_enabled: true
bam_enabled: true
soa_enabled: true
osb_enabled: false
repository_prefix: CDP
tag: mwconfig
The text was updated successfully, but these errors were encountered: