Attaching RBD device after qemu upgrade fails with "Unknown protocol 'rbd'"

Bug #1852766 reported by Louis Bouchard
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
qemu (Ubuntu)
New
Undecided
Unassigned
Bionic
Confirmed
Medium
Unassigned

Bug Description

After upgrading the following QEMU packages to latest Bionic version (1:2.11+dfsg-1ubuntu7.20) :

- qemu-system-x86
- qemu-system-common
- qemu-utils
- qemu-block-extra:amd64

Attaching an RBD device to a VM that was running prior to the upgrade fails with

"Unknown protocol 'rbd'

This only happens if this is the first RBD device to be attached after the upgrade. Attaching a device, then upgrading to 1:2.11+dfsg-1ubuntu7.20 will not generate the error message and the device will attach correctly.

Attaching a device to a new VM created _AFTER_ the upgrade does not trigger the problem either.

Downgrading qemu-block-extra:amd64 after the upgrade and prior to first attaching an RBD device will also not trigger the error.

Here are two examples :
Using :
ii qemu-block-extra:amd64 1:2.11+dfsg-1ubuntu7
ii qemu-system-common 1:2.11+dfsg-1ubuntu7
ii qemu-system-x86 1:2.11+dfsg-1ubuntu7
ii qemu-utils 1:2.11+dfsg-1ubuntu7
./qmp-shell -H /opt/vms/1/qmp.sock
(QEMU) drive_add auto format=raw,cache=writeback,file=rbd:rbd/disk-086264fd-9ca4-42a9-9deb-4a8be80f2bab:rbd_cache=true:id=compute-hv:conf=/etc/ceph/ceph.conf:keyring=/etc/ceph/ceph.client.compute-hv.keyring,id=volume-8b84f816-6e11-4a17-a2ce-0cdb6f486bfe,throttling.iops-total=10000,if=none
OK

With :
ii qemu-block-extra:amd64 1:2.11+dfsg-1ubuntu7
ii qemu-system-common 1:2.11+dfsg-1ubuntu7
ii qemu-system-x86 1:2.11+dfsg-1ubuntu7
ii qemu-utils 1:2.11+dfsg-1ubuntu7

and an existing QEMU process running

$ apt udpate && apt -y install qemu-utils qemu-system-x86 qemu-system-common qemu-block-extra:amd64
Using :
ii qemu-block-extra:amd64 1:2.11+dfsg-1ubuntu7.20
ii qemu-system-common 1:2.11+dfsg-1ubuntu7.20
ii qemu-system-x86 1:2.11+dfsg-1ubuntu7.20
ii qemu-utils 1:2.11+dfsg-1ubuntu7.20

./qmp-shell -H /opt/vms/1/qmp.sock
(QEMU) drive_add auto format=raw,cache=writeback,file=rbd:rbd/disk-086264fd-9ca4-42a9-9deb-4a8be80f2bab:rbd_cache=true:id=compute-hv:conf=/etc/ceph/ceph.conf:keyring=/etc/ceph/ceph.client.compute-hv.keyring,id=volume-8b84f816-6e11-4a17-a2ce-0cdb6f486bfe,throttling.iops-total=10000,if=none
Unknown protocol 'rbd'
(QEMU)

Restarting the QEMU process fixes the issue.

Louis Bouchard (louis)
Changed in qemu (Ubuntu Bionic):
status: New → Confirmed
importance: Undecided → Medium
description: updated
Revision history for this message
Christian Ehrhardt  (paelzer) wrote :

Hi Louis,
this is a known, but complex and probably very (very) long running issue.
I recently discussed it on the KVM Forum and we had a mail thread about it.

Read the full background on bug 1847361 - I'll mark this as a dup.

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.