Cinder-backup fails when cinder version isn't caught up to the rest of the deployment
Bug #1619004 reported by
Nate Potter
This bug affects 1 person
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Cinder |
Won't Fix
|
Medium
|
Unassigned | ||
os-brick |
Invalid
|
Medium
|
Unassigned |
Bug Description
During testing of rolling upgrades, one of the scenarios I had was the cinder service starting at stable/mitaka and the rest of the OpenStack services running at master, with the cinder-backup service running using the swift driver. In this setup, trying to create a backup of a volume fails and produces a backup in error status.
Here is the error in the c-bak log:
Changed in cinder: | |
importance: | Undecided → Medium |
Changed in os-brick: | |
importance: | Undecided → Medium |
To post a comment you must log in.
Interesting! Are c-bak and c-vol on a single node? Can you try to reproduce that while having both Mitaka's and Newton's volume.filters files in /etc/cinder/ rootwrap. d? Simply put both in that directory as for example volume- mitaka. filters and volume- newton. filters - oslo.rootwrap will load both.