Description of problem:
- We lost glusterfs endpoins after restart active master controllers host + 10-15miutes even though we created service for the endpoints.
Version-Release number of selected component (if applicable):
- OCP 3.3
- Restart "active" master controllers and wait for 10-15minutes
Steps to Reproduce:
1. Create endpoint and service.
2. Shut down host which is running active master-controllers
3. Wait for 10 or 15minutes.
* Need to reboot "active" master controllers hosts in step-2.
* Need to wait for 10-15minutes to see the endpoints disappearance in step-3.
- In step-3, GlusterFS endpoints gone.
- In step-3, GlusterFS endoints persist.
# cat gluster-service.yaml
- port: 1
# cat gluster-endpoints.yaml
- ip: 192.168.122.21
- port: 1
- ip: 192.168.122.22
- port: 1
Since this issue is critical for the enterprise system and we don't see any workaround, please backport to OCP 3.3 when you fixed the issue.
Kenjiro, In above example, the service and endpoint are having different names. Is there a typo? if not, the headless service has to be connected with endpoint with the same name.
Hi, thank you for your update. What does "same name" means? Does it mean metadata.name? The configuration is exactly same with example in the official docs - https://docs.openshift.com/container-platform/3.3/install_config/storage_examples/gluster_example.html#complete-example-using-gusterfs-creating-the-gluster-endpoints
(In reply to Kenjiro Nakayama from comment #3)
> Hi, thank you for your update. What does "same name" means? Does it mean
Yes, that's right. The "metadata: name:" defined in your Gluster service definition must be defined in the endpoints definition as well to match the endpoints to this service. So in short, the "metadata: name:" should be the same in both Gluster service definition and Gluster endpoint definition.
The configuration is exactly same with example in the
> official docs -
The official doc is actually here: https://access.redhat.com/documentation/en/openshift-container-platform/3.3/paged/installation-and-configuration/chapter-19-configuring-persistent-storage#install-config-persistent-storage-persistent-storage-glusterfs
Please try using the examples mentioned in the above official doc and let us know if you still face any issues.
Hi Humble and Prasanth,
Thank you. I confirmed that same name could work fine. But actually, both docs are RH "official" docs, so we have to update the example one. Could you assign this ticket to the docs team? Or should I open new ticket?
Sent email to @jparra asking for additional information.
This was fixed via: