Hi Matthew,
I can confirm that there is something with kernel 6.0.18-200.fc36 (We
have it on our OKD 4 nodes).
When we upgraded OKD (Opnshift 4 upstream with Fedora CoreOS 36) to
4.11.0-0.okd-2023-01-14-152430, it upgraded the Kernel 6.0.10-200.fc36
-> 6.0.18-200.fc36.
We use Rook-CEPH operator for Kubernetes to provision our CEPH in OKD.
After that we could observe OSDs in "peering" state, randomly, no system.
We checked network connections between the CEPH Pods and it worked well,
no issue.
Thank you for your report, it gave us a hint, we downgraded the Kernel
back to 6.0.10-200.fc36 and CEPH started working again.
// Dmitrii.
_______________________________________________
ceph-users mailing list -- ceph-users@xxxxxxx
To unsubscribe send an email to ceph-users-leave@xxxxxxx
