This message was deleted Rancher Users #harvester

This message was deleted.

adamant-kite-43734

08/06/2026, 8:56 AM

This message was deleted.

bland-article-62755

08/06/2026, 2:33 PM

I can test on 1.7.1

bland-article-62755

08/06/2026, 2:53 PM

I couldn't reproduce it on 1.7.1

bland-article-62755

08/06/2026, 2:54 PM

granted it was GPUs and not disk, but maybe that's part of it?

bland-article-62755

08/06/2026, 2:55 PM

Is there anything in the pod logs before restart pci devices controller when it happens?

bland-article-62755

08/06/2026, 3:14 PM

Is it always the same host?

bland-article-62755

08/06/2026, 3:16 PM

I wonder if it could be a badly behaving pci card or drive

gorgeous-guitar-43479

08/07/2026, 2:12 PM

I tried it a couple of times, and I couldn't reproduce it again.

gorgeous-guitar-43479

08/07/2026, 2:13 PM

It looks like some race condition in informers in kubernetes operators ...

gorgeous-guitar-43479

08/07/2026, 2:14 PM

Maybe the PCI extension was scheduled to the node that has been evicted

2 Views