This message was deleted.
# longhorn-storagea
adamant-kite-43734
04/28/2024, 7:20 AMThis message was deleted.
n
nice-tent-65195
04/29/2024, 11:16 AMHello there!
Can you check the logs of the pre-upgrade pod?
nice-tent-65195
04/29/2024, 11:18 AMI have had a similar phenomen. The cause was for me was an old, stucked longhorne engine image. Can you show the output of
kubectl -n longhorn-system get lheic
crooked-cat-21365
04/29/2024, 11:52 AM Copy code
{hdunkel@dpcl082:~ 1001} kubectl -n longhorn-system get lhei
NAME INCOMPATIBLE STATE IMAGE REFCOUNT BUILDDATE AGE
ei-881c42db deployed rancher/mirrored-longhornio-longhorn-engine:v1.3.2 18 563d 452d
ei-8c052ab7 deployed rancher/mirrored-longhornio-longhorn-engine:v1.5.3 48 164d 146dn
nice-tent-65195
04/29/2024, 11:54 AMYep, then same issue for you too. I assume that you're on 1.5.3 LH...
c
crooked-cat-21365
04/29/2024, 11:55 AMThe other version is 1.3.2
n
nice-tent-65195
04/29/2024, 11:57 AMYou should go to the Longhorn UI/Volume tab and upgrade the volumes that are using the old engine:
https://longhorn.io/docs/1.5.3/deploy/upgrade/upgrade-engine/#live-upgrade
After that, Longhorn will remove the old 'lhei' object, if it doesn't then you should follow this: https://longhorn.io/docs/1.5.3/deploy/upgrade/upgrade-engine/#clean-up-the-old-image
c
crooked-cat-21365
04/29/2024, 11:58 AMI just wonder why my other kubes didn't run into this problem?
n
nice-tent-65195
04/29/2024, 12:01 PMI assume you don't have old lhei in those clusters. I don't know why it's stucked here.
c
crooked-cat-21365
04/29/2024, 12:04 PMI have to talk with my colleagues to clean up detached and forgotten volumes first. Stay tuned.
👍 1
crooked-cat-21365
04/29/2024, 12:44 PMThere were 6 detached PVs (each 411 days old). The owner confirmed that I can remove the entire namespace. After that the problem was gone.
I highly appreciate your help on this.
👍 1
104 Views