- Incident summary/Résumé de l'incident: Errors during volumes attachment to pods
- Start time/Heure de début: 2020-05-19 03:01:09.931 UTC
- Impact / Périmètre affecté: All Managed Kubernetes Service regions
- Impact type / Type d'impact : Creation and rescheduling of pods with volumes
- Estimated date to recovery / Date de résolution estimée : NO ETA to share at the moment
* No issue on APIServer and others available resources.
* No issue on persistent volumes created on PublicCloud side.
Update(s):
Date: 2020-05-22 16:52:29 UTC The incident is now solved.
Pods can now mount more than 6 volumes by nodes available in managed clusters.
Date: 2020-05-22 16:47:24 UTC Patch deployment on BSH5 is now over.
Date: 2020-05-22 16:23:43 UTC Patch deployment on GRA5 is now over.
Patch deployment on BHS5 in progress.
Date: 2020-05-22 15:46:59 UTC Patch deployment on GRA7 is now over.
Patch deployment on GRA5 in progress.
Date: 2020-05-22 14:55:40 UTC Patch deployment in progress on GRA7 region.
Date: 2020-05-22 13:31:50 UTC We are still working with the Public CLoud team to fix this outage.
Sorry for the inconvenience.
Date: 2020-05-22 10:14:46 UTC The root cause has been found in the libvirt toolkit used by Openstack.
No more than 6 volumes can be mounted per host.
A patch will be implemented and deployed in the next hours.