Hide Forgot
Description of problem: When a persistent volume is requested for shared access (RWX), the request remains indefinitely pending. How reproducible: Everytime Steps to Reproduce: 1. On a project, click Request storage 2. Enter a name: eg. "stroage", choose"Shared Access (RWX)" for "Access mode", and capacity of 1GiB. 3. Click create and check storage creation progess. Eg. https://console.preview.openshift.com/console/project/[projectname]/browse/storage Actual results: Creation remains in pending mode for a very long time (indefinitely?) storage Pending - RWX (Read-Write-Many) 2 days Expected results: Creation to succeed in a reasonable time. storage Bound to volume pv-1-preview-master-afbb8-vol-02ec3568e7a0ab265 1 GiB RWX (Read-Write-Many) 2 days Additional info: Quotas are taken into consideration. 1 PVC exists and maximum size of 1GiB.
There was a transient issue related to persistent volumes over the weekend and that issue was resolved on Monday. Does your issue still persist?
The issue is still exist, and a pv with ROX accessmode can not be requested too. 1. Request a RWX pv from web console: mongodb-2 Pending - RWX (Read-Write-Many) 11 minutes 2. Request a ROX pv from web console: mongodb-3 Pending - ROX (Read-Only-Many) 11 minutes
We must validate the AccessModes on the PVs that are precreated and managed by Ops are correct. The PVC would also require matching AccessModes. This was not an issue when we were dynamically provisioning EBS volumes. The accessModes were ignored. But since we are managing a manually provisioned pool of volumes and relying on matching, the modes must match. @yasun, can you test this with the correct modes (RXW for EBS volumes)?
Can successfully request a pv with RWO accessmode. @Mark, I am not sure whether the correct mode "RXW" you mentioned at comment#5 are correct. I guess it should be "RWO". so I do with the three accessmodes again: 1. Request a RWO pv from web console: yasun-test-3 Bound to volume pv-1-preview-master-afbb8-vol-0db53afa77c3354d6 1 GiB RWO (Read-Write-Once) a few seconds 2. Request a ROX pv from web console: yasun-test-2 Pending - ROX (Read-Only-Many) 22 minutes 3. Request a RWX pv from web console: yasun-test-1 Pending - RWX (Read-Write-Many) 24 minutes
Same here, setting up: * RWO (Read-Write-Once) is near instant * RWX (Read-Write-Many) is still pending
@yasun, you are correct. I made a typo with RWX, I meant RWO. If the manually provisioned pool of storage only contains RWO volumes, you should only expect PVCs w/ RWO to match. Any others will remain Pending indefinitely because there are no volumes that match what the user requested.
I have added some RWX volumes to the pool to work around the issue until we get auto-provisioning back. RWX PVCs should work now.
> RWX PVCs should work now. *Matching* those PVs will work, but that behavior will not. EBS volumes are RWO.
The RWX pvc can work: 1. Request a RWX pv from web console: yasun-test-1 Bound to volume pv-1-preview-master-afbb8-vol-06fefdbacccfcfc93 1 GiB RWX (Read-Write-Many) 17 minutes @Gerard Braad, has the work around fit for your requirement?
@yasun, the PVC completed successfully. However I have been unable to test anything, as the account got disabled/expired.
@Gerard Braad, how about the progress in your site? have you apply the access and to do test on https://console.preview.openshift.com/console?
@yasun I have applied but hasn't been recreated. But according to my previous feedback it currently works as expected. And therefore the bug can be closed. I do believe a different issue would have to be created related to the PVC pending for a long time. Somehow a timeout should trigger a notification as a warning or error.