Bug 1849488 - [OSP 13] "Failed to attach volume at mountpoint" error when adding a volume to a fresh RHEL 8.2 PPC VM
Summary: [OSP 13] "Failed to attach volume at mountpoint" error when adding a volume t...
Keywords:
Status: CLOSED ERRATA
Alias: None
Product: Red Hat OpenStack
Classification: Red Hat
Component: openstack-nova
Version: 13.0 (Queens)
Hardware: ppc64le
OS: Linux
high
high
Target Milestone: z13
: 13.0 (Queens)
Assignee: Lee Yarwood
QA Contact: Paras Babbar
URL:
Whiteboard:
Depends On:
Blocks: 1838786
TreeView+ depends on / blocked
 
Reported: 2020-06-22 04:36 UTC by ldenny
Modified: 2024-06-13 22:47 UTC (History)
19 users (show)

Fixed In Version: openstack-nova-17.0.13-14.el7ost
Doc Type: If docs needed, set a value
Doc Text:
Clone Of:
Environment:
Last Closed: 2020-10-28 18:32:10 UTC
Target Upstream Version:
Embargoed:


Attachments (Terms of Use)


Links
System ID Private Priority Status Summary Last Updated
IBM Linux Technology Center 186566 0 None None None 2020-06-30 08:24:18 UTC
OpenStack gerrit 730313 0 None MERGED fix scsi disk unit number of the attaching volume when cdrom bus is scsi 2020-12-16 22:01:47 UTC
Red Hat Issue Tracker OSP-32277 0 None None None 2024-06-13 22:47:08 UTC
Red Hat Product Errata RHBA-2020:4393 0 None None None 2020-10-28 18:32:25 UTC

Description ldenny 2020-06-22 04:36:10 UTC
Description of problem:
When trying to attach a volume to a fresh  RHEL 8.2 PPC VM the following error is seen in the nova logs:

~~~
Failed to attach volume at mountpoint: /dev/sdc: libvirtError: Requested operation is not valid: Domain already contains a disk with that address
~~~

Version running on the compute nodes: libvirt-4.5.0-23.el7_7.5.x86_64

Comment 7 David Gibson 2020-06-25 02:23:49 UTC
I'm confused.  The title says this is a ppc VM, but then an x86 libvirt package is listed as the version.  The comments say we're dealing with RHEL8.2, but it's filed against RHEL7.7.

Can we get some more clarity on the situation here:
   1. What's the Openstack version in use?
   2. What's the RHEL version on the hosts?
   3. What type of machines are the compute hosts?
   4. What OS is in the guests?
   5. What are package versions for more relevant components:
       - host kernel
       - guest kernel
       - Nova
       - qemu-kvm

virtio-scsi should definitely work with ppc now, although it's possible if one of the components is an old version it might lack support.

The initial error suggests that Nova might be instructing libvirt to connect two disks at the same address, which sounds like a Nova or configuration bug, rather than a libvirt or

Comment 8 ldenny 2020-06-25 02:37:25 UTC
Hi David,

Sorry, the libvirt version running on the compute node is libvirt-4.5.0-23.el7_7.5.ppc64le

I am just gathering the other information for you.

Comment 9 ldenny 2020-06-25 07:07:55 UTC
1. What's the OpenStack version in use? 
OSP13z11

2. What's the RHEL version on the hosts? 
Red Hat Enterprise Linux Server release 7.7 (Maipo)

3. What type of machines are the compute hosts i.e brand and model?
IBM Power 822L

4. What OS is in the guests?
RHEL8.2 PPC but the issue is the same for RHEL7.6, RHEL8.2 and Fedora28

5. What are the package versions for more relevant components:
    - host kernel
Linux compute-dev-822l-0 3.10.0-1062.12.1.el7.ppc64le #1 SMP Thu Dec 12 11:47:54 UTC 2019 ppc64le ppc64le ppc64le GNU/Linux

    - guest kernel
Different on RHEL7.6, RHEL8.2 and Fedora28

    - Nova
openstack-nova-compute:13.0-129

    - qemu-kvm 
libvirt-daemon-driver-qemu-4.5.0-23.el7_7.5.ppc64le
qemu-kvm-rhev-2.12.0-33.el7_7.4.ppc64le
qemu-kvm-common-rhev-2.12.0-33.el7_7.4.ppc64le

Let me know if you need anything else.

Comment 10 Jaroslav Suchanek 2020-06-26 10:17:04 UTC
Hi David, thank you for stepping in. Please review the info in comment 9. Is there anything else you need to know? Thanks.

Comment 11 Laurent Vivier 2020-06-26 12:20:39 UTC
It looks like a known problem:
https://access.redhat.com/solutions/4356171

Could you check?

Comment 12 Laurent Vivier 2020-06-26 12:28:50 UTC
(In reply to Jaroslav Suchanek from comment #10)
> Hi David, thank you for stepping in. Please review the info in comment 9. Is
> there anything else you need to know? Thanks.

Can someone from libvirt team check if the problem described here is the same as the one described in https://access.redhat.com/solutions/4356171 ?

Comment 13 Daniel Berrangé 2020-06-26 12:47:33 UTC
(In reply to Laurent Vivier from comment #12)
> (In reply to Jaroslav Suchanek from comment #10)
> > Hi David, thank you for stepping in. Please review the info in comment 9. Is
> > there anything else you need to know? Thanks.
> 
> Can someone from libvirt team check if the problem described here is the
> same as the one described in https://access.redhat.com/solutions/4356171 ?

It doesn't look like it.

That page says the problem was fixed in libvirt-4.5.0-23.el7_7.3.  Comment #9 shows the customer already has 4.5.0-23.el7_7.5.ppc64le, so will have the fix.

In addition the bug on that page hits when the VM has more than 6 disks, and attaching a disk on unit=7

This VM only has 2 disks and is requesting sdc which should get unit=2.

I think we'd need to see the XML that the Nova is using to hotplug the disk, to identify whether the info it provides to libvirt is correct.

Comment 16 IBM Bug Proxy 2020-06-30 08:31:36 UTC
------- Comment From chavez.com 2020-06-29 15:40 EDT-------
> Comment #9 shows the customer already has 4.5.0-23.el7_7.5.ppc64le, so will have the fix.

So can you confirm this is bug is for a problem reported by an external customer please? Thanks.

Comment 38 errata-xmlrpc 2020-10-28 18:32:10 UTC
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory (openstack-nova bug fix advisory), and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

https://access.redhat.com/errata/RHBA-2020:4393


Note You need to log in before you can comment on or make changes to this bug.