Cinder LVM over iSCSI driver under qemu 2.0 and kernel <3.11
Bug #1375245 reported by
Pavel Boldin
This bug affects 1 person
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Mirantis OpenStack |
Fix Released
|
High
|
Pavel Boldin | ||
5.1.x |
Fix Committed
|
High
|
Pavel Boldin | ||
6.0.x |
Won't Fix
|
High
|
Pavel Boldin | ||
6.1.x |
Fix Released
|
High
|
Pavel Boldin |
Bug Description
Usage of the Cinder LVM over iSCSI in Fuel 5.1.1 leads to messages of hanged qemu-system-x86_64 process being produced by the kernel.
This is caused by the call 'fdatasync' called by the qemu on the iSCSI device when the kernel inside qemu emits 'SYNCHRONIZE_CACHE' SCSI command. The kernel code that handles this route (blkdev_
The solution is either to update to a newer kernel (3.12) that has a completely new block queue implementation or to update qemu not to try to sync entire iSCSI partition at once.
Changed in mos: | |
milestone: | none → 6.0 |
tags: | added: cinder |
Changed in mos: | |
status: | New → Triaged |
Changed in mos: | |
status: | In Progress → Fix Committed |
status: | Fix Committed → In Progress |
Changed in mos: | |
milestone: | 6.0 → 6.1 |
no longer affects: | mos/6.1.x |
Changed in mos: | |
status: | In Progress → Confirmed |
status: | Confirmed → Triaged |
status: | Triaged → In Progress |
status: | In Progress → Confirmed |
status: | Confirmed → In Progress |
milestone: | 6.1 → 6.0.1 |
status: | In Progress → Won't Fix |
To post a comment you must log in.
Using a novel kernel (3.13.0-36-generic) does not help to solve this problem.
The only other solution is to flush a device step by step. For this to be implemented `qemu` and `tgtd` must be updated.