OpenShift Container Storage 3.11.3 : Resolved Issues and Enhancements
Updated
OpenShift Container Storage 3.11.3 is a batch release for OpenShift Container Storage 3.11 and can be applied as an update to an existing installation or can be used to create a new installation. Note, OpenShift Container Storage 3.11 updates are cumulative and include fixes from previous releases.
OpenShift Container Storage 3.11.3 Errata
The following list provides errata for OpenShift Container Storage 3.11.3 release:
- cns-deploy-tool bug fix update
- rhgs-gluster-block-prov-container bug fix update
- rhgs-volmanager-container bug fix update
- rhgs-server-container bug fix update
- rhgs-s3-server-container bug fix update
- heketi bug fix and enhancement update
- gluster-block bug fix update
Notable Enhancements
Add option to limit the number of heketi volumes which can be created on heketi cluster(BZ#1476223)
Heketi now enforces a limit on the number of file volumes that can be
created in a Gluster trusted storage pool. The default limit is set to 1000
volumes. To override this value, admins should set the
max_volumes_per_cluster option in 'glusterfs' section of heketi.json or set
the env variable HEKETI_GLUSTER_MAX_VOLUMES_PER_CLUSTER.
Provide a way for heketi to start in an "administrative mode" (BZ#1691509)
Previous versions of Heketi allowed setting an "administrative mode" using the heketi-cli tool. These modes could be used to prepare the server for shut down or restrict access to locally issued commands. However, these modes could only be applied after the server was started. In some circumstances, allowing the server to start normally before the administrative mode was set opened a window for undesired changes to be requested by clients. Now, an initial state at startup via the HEKETI_DEFAULT_STATE environment variable can be provided to Heketi.
Known Issues
In app pods, mpath device name is not mapped/created for some block volumes in the initiator side (BZ#1671364)
When an application pod is created with multipath configuration and a gluster-block PVC, the iSCSI driver waits for 10 seconds for the storage path discovery. However, device discovery can be delayed when the system is under load or similar conditions. When only one path is discovered within 10 seconds, this path is used to mount the volume to the application pod. This means that failover to a different path does not work even when multiple paths were configured. To work around this issue, delete and reschedule the pod; there is a high possibility that, the new pod mounts the volume with all the paths.
Mismatch of heketi-cli and gluster volume list. after re installation of ocs . gluster has volumes which were there before uninstall. (BZ#1700662)
Uninstallation of the OCS playbook fails to clean the /var/lib/glusterd directory on the nodes where glusterfs pods are deployed. As a consequence, if a new OCS is deployed with *at least* one glusterfs pod on the old nodes, then the volumes created in the previous deployment are visible in the gluster command outputs. However, these volumes are seen neither in OpenShift nor in Heketi thereby causing confusion. This has no impact on the dynamic provisioning of the PVCs and usage of PVs created in the new deployment. However, device/node replace operations of Heketi might fail. To workaround this issue, you need to identify the list of volumes that were created in the previous deployment and delete the relevant directories and files from /var/lib/glusterd/ and /var/lib/heketi/fstab.
Bugs Fixed in OpenShift Container Storage 3.11.3
| Bug ID | Summary |
|---|---|
| This content is not included.1701912 | rebuild cns-deploy for 3.11.3 |
| This content is not included.1701915 | rebuild rhgs-gluster-block-prov-container for 3.11.3 |
| This content is not included.1711752 | Respin rhgs-gluster-block-prov container for OCS 3.11.3 with signed rpms |
| This content is not included.1686894 | Heketi container should backup db before heketi process is started |
| This content is not included.1701916 | Rebuild rhgs-volmanager-container for 3.11.3 |
| This content is not included.1711748 | Respin rhgs-volmanager container for OCS 3.11.3 with signed rpms |
| This content is not included.1676466 | LVM in the glusterfs container should not try to use udev |
| This content is not included.1685935 | disable dmeventd in the rhgs-server container |
| This content is not included.1700814 | Add GB_CLI_TIMEOUT env configuration in rhgs-server-container template/container scripts. |
| This content is not included.1707340 | respin rhgs-server container to include latest gluster-block packages |
| This content is not included.1708577 | fresh install gluster pod not in 1/1 state tcmu-runner failed to find var/run/lock/tcmu.lock |
| This content is not included.1711750 | Respin rhgs-server container for OCS 3.11.3 with signed rpms |
| This content is not included.1476223 | [RFE ] add option to limit number of heketi volumes which can be created on heketi cluster |
| This content is not included.1685005 | Heketi sets "user.heketi.zone-checking" to strict or none based on DC/template settings & not based on settings in the SC used |
| This content is not included.1691509 | [RFE] Provide a way for heketi to start in an "administrative mode" |
| This content is not included.1701990 | Respin of rhgs-s3-server container for OCS-3.11.3 |
| This content is not included.1707843 | Respin of RHGS 3 S3 Server Container image container to include CVE fixed at python-jinja2 packages |
| This content is not included.1711751 | Respin rhgs-s3-server container for OCS 3.11.3 with signed rpms |
| This content is not included.1661406 | gluster-blockd: There is no notification on changing the log-level value to ERROR and WARNING in gluster-blockd config file |
| This content is not included.1665543 | [GSS] Provisioning 2TB block volume results in timeout |
Product(s)
Category
Article Type