Skip to content

HTTPS clone URL

Subversion checkout URL

You can clone with
or
.
Download ZIP
Commits on Jul 28, 2015
  1. @davidvossel

    Merge pull request #648 from guessi/fix_docker_images_lookup

    davidvossel authored
    Fix docker images lookup
  2. @dmuhamedagic

    Use runuser inteald of su command for SELinux enforcing mode

    HashimoyoNaoya authored dmuhamedagic committed
  3. @guessi
  4. @guessi

    remove unused variable

    guessi authored
Commits on Jul 27, 2015
  1. @guessi

    Fix "docker images" difference between distro

    guessi authored
    On Ubuntu:
    
        $ sudo docker images
    
        REPOSITORY          TAG
        ubuntu              latest
    
    On Fedora:
    
        $ sudo docker images
    
        REPOSITORY          TAG
        docker.io/ubuntu    latest
  2. @guessi
  3. @pemer @fghaas

    iSCSITarget: properly create portals for lio-t implementation

    pemer authored fghaas committed
    Even when defining the "portals" parameter, targetcli creates a
    default portal at 0.0.0.0:3260 if the auto_add_default_portal option
    is set to true (default), causing the resource to fail with
    OCF_ERR_GENERIC.
    
    Suppress creation of the default portal by setting
    auto_add_default_portal=false if the "portals" parameter is set,
    and retain the original behavior if it is not.
    
    Fixes ClusterLabs/resource-agents#630.
  4. @fghaas

    iSCSILogicalUnit: when deleting a LUN or initiator fails with lio-t, …

    fghaas authored
    …proceed with warning
    
    During the stop operation, "targetcli delete" returns an error if the
    object we're about to delete already no longer exists. Since stop is a
    force operation and we must do anything we can to clean up the
    resource, don't exit on that condition but log a warning and proceed
    to removal of the backstore. If *that* fails, then we error out and
    potentially cause the node to be fenced.
    
    Fixes ClusterLabs/resource-agents#642.
  5. @fghaas

    iSCSILogicalUnit: return OCF_NOT_RUNNING on monitor if backing path d…

    fghaas authored
    …oes not exist
    
    If iSCSILogicalUnit exposes a local LVM logical volume or even a
    hardware drive, then it is safe to assume that the device will
    "always" be there. However, if what we are exposing is itself a
    cluster resource, such as an RBD volume or a DRBD device, then the
    backing device will appear and disappear as needed.
    
    It is safe to assume that if the backing file or block device does not
    exist, then it cannot be exported as an iSCSILogicalUnit. Even if it
    *has* been previously exported as an iSCSILogicalUnit, then it should
    be impossible for it to just disappear or be removed while open, as
    such an action would return -EBUSY while the device is open. If it
    disappears because the drive has physically been yanked, then that is
    a condition in which throwing up our hands and fencing the node is the
    proper action.
    
    Thus, if during monitor the backing device does not exist, return
    OCF_NOT_RUNNING.
    
    Fixes ClusterLabs/resource-agents#641.
  6. @fghaas

    iSCSILogicalUnit: add check for leftover target/core entries for lio-t

    fghaas authored
    Apply the same checking behavior to the lio-t target as the RA
    previously already did for lio: do not report the resource as being
    gracefully stopped if it still has iblock handles open.
    
    With lio-t, we cannot specify the iblock number, but since we're using
    $OCF_RESOURCE_INSTANCE as the backstore key here (which is unique
    across the cluster), the more inclusive file glob is still safe.
    
    Fixes ClusterLabs/resource-agents#640.
Commits on Jul 22, 2015
  1. @dmuhamedagic

    Merge pull request #646 from ytakeshita/fix_initialize_replication_slot3

    dmuhamedagic authored
    Medium: pgsql: delete an old replication slot when creating the slot.
  2. @ytakeshita

    Medium: pgsql: delete an old replication slot when creating the slot.

    ytakeshita authored
    Initialize the replication slot (delete an old slot if exists and create a new one)
    to prevent that the old slot runs out of disk space.
    
    Without the initialization, PostgreSQL would try to use the old slot after a node
    failed once and re-joined to the cluster, and continue to accumulate WAL files for
    the old slot, which would eventually run out of disk space.
  3. @ytakeshita
Commits on Jul 20, 2015
  1. @dmuhamedagic

    Merge pull request #645 from sathieu/patch-1

    dmuhamedagic authored
    Support RozoFS
Commits on Jul 17, 2015
  1. @sathieu

    Support RozoFS

    sathieu authored
Commits on Jun 30, 2015
  1. @davidvossel

    Merge pull request #635 from ctowsley1/orainstance-rgmanager-returnvalue

    davidvossel authored
    orainstance.sh: interpret listener stop results correctly
Commits on Jun 29, 2015
  1. @davidvossel

    Merge pull request #627 from davidvossel/dhcpd-default

    davidvossel authored
    High: dhcpd: use correct default chroot for rhel based systems
Commits on Jun 25, 2015
  1. @davidvossel
Commits on Jun 15, 2015
  1. @davidvossel

    Merge pull request #621 from guessi/fix_docker_existance_test

    davidvossel authored
    Fix docker existance test
Commits on Jun 13, 2015
  1. @guessi

    Docker: fix docker existance test report error

    guessi authored
    according to `docker inspect --help`, `docker inspect` may report information
    for both container and image, if there's a image named "XYZ", and trying to
    inspect "XYZ" w/o specifying `--format {{.A.B.C}}`, it would report image
    state instead of container state.
    
    $ sudo docker inspect --help
    
        Usage: docker inspect [OPTIONS] CONTAINER|IMAGE [CONTAINER|IMAGE...]
    
        Return low-level information on a container or image
    
    and further more, we should judge if it is really a container,
    here we use `--format {{.State.Running}}`, and more, we should judge if
    it is really a container by determining return value, if it is a container,
    we should get a string of boolean value "true" or "false", and we can tell it
    is a Docker container, but not a Docker image.
    
    if it is a Docker image,
    $ sudo docker inspect --format {{.State.Running}} ubuntu:latest
        <no value>
    
    if it is a running Docker container,
    $ sudo docker inspect --format {{.State.Running}} my_container
        true
    
    if it is a stopped Docker container,
    $ sudo docker inspect --format {{.State.Running}} my_container
        false
  2. @guessi

    Docker: remove trailing spaces

    guessi authored
Commits on Jun 4, 2015
  1. @davidvossel

    Merge pull request #617 from ytakeshita/fix-replication-slot

    davidvossel authored
    Low: pgsql: eliminate duplicate "ocf_exit_reason".
Commits on Jun 2, 2015
  1. @ytakeshita
Commits on May 22, 2015
  1. @dmuhamedagic

    Merge pull request #614 from krig/travis-ci

    dmuhamedagic authored
    Dev: Add Travis CI job which runs shellcheck on all scripts
Commits on May 19, 2015
  1. @krig

    Dev: Add Travis CI job which runs shellcheck on all scripts

    krig authored
    This adds a Travis CI which invokes shellcheck on all scripts
    in the repository. It doesn't currently signal failure, but the
    build job output can be inspected for any issues that should
    be fixed.
Commits on May 15, 2015
  1. @davidvossel

    Merge pull request #604 from davidvossel/clvmd-activate

    davidvossel authored
    High: clvm: activate_vgs option for enable/disable of automatic vg activ...
  2. @davidvossel

    Merge pull request #605 from davidvossel/ipsrcaddr_stop

    davidvossel authored
    High: IPsrcaddr: return correct error code during stop when misconfigure...
  3. @davidvossel

    Merge pull request #613 from elconas/master

    davidvossel authored
    Fixed redis start operation if replication sync takes > 20 seconds
Commits on May 9, 2015
  1. Made 'slave of XXX' wait forever (demote)

    Robert Heinzmann authored
Commits on May 8, 2015
Commits on May 6, 2015
  1. @davidvossel

    Merge pull request #608 from davidvossel/ethmon-simple

    davidvossel authored
    High: ethmonitor: link_status_only option for skipping RX counter and ar...
Commits on May 1, 2015
  1. @davidvossel

    Merge pull request #603 from davidvossel/galera-fixup

    davidvossel authored
    Minor galera fixes
Commits on Apr 29, 2015
  1. @davidvossel
Commits on Apr 28, 2015
  1. @davidvossel
Something went wrong with that request. Please try again.