Volume initialisation with image might fail due to disk caching
Bug #1363016 reported by
Dirk Mueller
This bug affects 1 person
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Cinder |
Fix Released
|
Undecided
|
Dirk Mueller |
Bug Description
When using the volume initialisation with an image, not all of the image might be written back on the volume before it gets unmapped due to not flushing/avoiding disk caching on write. The problem seems to occur mostly with volumes that are provided via multipath/Fibre Channel, e.g. where the LUNs disappear without the cinder-volume guest node having full control over it (and the kernel there being able to trigger a flush in time).
Changed in cinder: | |
assignee: | nobody → Dirk Mueller (dmllr) |
status: | New → In Progress |
Changed in cinder: | |
milestone: | none → juno-rc1 |
status: | Fix Committed → Fix Released |
Changed in cinder: | |
milestone: | juno-rc1 → 2014.2 |
no longer affects: | ubuntu |
tags: | added: icehouse-backport-potential |
To post a comment you must log in.
Reviewed: https:/ /review. openstack. org/112033 /git.openstack. org/cgit/ openstack/ cinder/ commit/ ?id=5c5f3c0684c aada0fb7efe33d5 9bc7f1d66c72d8
Committed: https:/
Submitter: Jenkins
Branch: master
commit 5c5f3c0684caada 0fb7efe33d59bc7 f1d66c72d8
Author: Dirk Mueller <email address hidden>
Date: Tue Aug 5 16:13:33 2014 +0200
Avoid using the disk cache on volume initialisation
When caching is involved, the volume might be
unmapped before the copy actually hit the disk or the
VM starts booting before all of the data has been flushed,
which causes the VM to crash at an arbitrary point in time.
Closes-Bug: #1363016
Change-Id: I7a04f683add8c2 3b9125fe837c404 8ccc3ac224d