8:33 a.m.
On Tue, Aug 14, 2018 at 6:03 PM Алексей Максимов <
aleksey.i.maksimov(a)yandex.ru> wrote:
Hello, Nir
Log in attachment.
In the log we can see both createVolume and deleteVolume fail for this disk
uuid:
cdf1751b-64d3-42bc-b9ef-b0174c7ea068
1. Please share the output of this command on one of the hosts:
lvs -o vg_name,lv_name,tags | grep cdf1751b-64d3-42bc-b9ef-b0174c7ea068
This will show all the volumes belonging to this disk.
2. For every volume, share the output of qemu-img info:
If the lv is not active, activate it:
lvchange -ay vg_name/lv_name
Then run qemu-img info to find the actual chain:
qemu-img info --backing /dev/vg_name/lv_name
If the lv was not active, deactivate it - we don't want to leave unused lvs
active.
lvchange -an vg_name/lv_name
3. On of the volume in the chain will not be part of the chain.
No other volume will use it as backing file, and it may not have a backing
file, or it may point to another volume in the chain.
Once we found this volume, please check engine logs for this volume uuid.
You will probaly
find that the volume was deleted in the past. Maybe you will not find it
since it was deleted
months or years ago.
4. To verify that this volume does not have metadata, check the volume MD_N
tag.
N is the offset in 512 bytes blocks from the start of the metadata volume.
This will read the volume metadata block:
dd if=dev/vg_name/metadata bs=512 count=1 skip=N iflag=direct
We expect to see:
NONE=####################################################...
5. To remove this volume use:
lvremove vg_name/lv_name
Once the volume is removed, you will be able to create snapshot.
14.08.2018, 01:30, "Nir Soffer" <nsoffer(a)redhat.com>:
On Mon, Aug 13, 2018 at 1:45 PM Aleksey Maksimov <
aleksey.i.maksimov(a)yandex.ru> wrote:
We use oVirt 4.2.5.2-1.el7 (Hosted engine / 4 hosts in cluster / about
twenty virtual machines)
Virtual machine disks are located on the Data Domain from FC SAN.
Snapshots of all virtual machines are created normally. But for one
virtual machine, we can not create a snapshot.
When we try to create a snapshot in the oVirt web console, we see such
errors:
Aug 13, 2018, 1:05:06 PM Failed to complete snapshot 'KOM-APP14_BACKUP01'
creation for VM 'KOM-APP14'.
Aug 13, 2018, 1:05:01 PM VDSM KOM-VM14 command HSMGetAllTasksStatusesVDS
failed: Could not acquire resource. Probably resource factory threw an
exception.: ()
Aug 13, 2018, 1:05:00 PM Snapshot 'KOM-APP14_BACKUP01' creation for VM
'KOM-APP14' was initiated by petya@sub.holding.com(a)sub.holding.com-authz.
At this time on the server with the role of "SPM" in the vdsm.log we see
this:
...
2018-08-13 05:05:06,471-0500 INFO (jsonrpc/0) [jsonrpc.JsonRpcServer]
RPC call VM.getStats succeeded in 0.00 seconds (__init__:573)
2018-08-13 05:05:06,478-0500 INFO (jsonrpc/5) [jsonrpc.JsonRpcServer]
RPC call Image.deleteVolumes succeeded in 0.05 seconds (__init__:573)
2018-08-13 05:05:06,478-0500 INFO (tasks/3)
[storage.ThreadPool.WorkerThread] START task
bb45ae7e-77e9-4fec-9ee2-8e1f0ad3d589 (cmd=<bound method Task.commit of
<vdsm.storage.task.Task instance at 0x7f06b85a2128>>, args=None)
(threadPool:208)
2018-08-13 05:05:07,009-0500 WARN (tasks/3) [storage.ResourceManager]
Resource factory failed to create resource '01_img_6db73566-0f7f-4438-a9ef-
6815075f45ea.cdf1751b-64d3-42bc-b9ef-b0174c7ea068'. Canceling request.
(resourceManager:543)
Traceback (most recent call last):
File "/usr/lib/python2.7/site-packages/vdsm/storage/resourceManager.py",
line 539, in registerResource
obj = namespaceObj.factory.createResource(name, lockType)
File
"/usr/lib/python2.7/site-packages/vdsm/storage/resourceFactories.py", line
193, in createResource
lockType)
File
"/usr/lib/python2.7/site-packages/vdsm/storage/resourceFactories.py", line
122, in __getResourceCandidatesList
imgUUID=resourceName)
File "/usr/lib/python2.7/site-packages/vdsm/storage/image.py", line 213,
in getChain
if srcVol.isLeaf():
File "/usr/lib/python2.7/site-packages/vdsm/storage/volume.py", line
1430, in isLeaf
return self._manifest.isLeaf()
File "/usr/lib/python2.7/site-packages/vdsm/storage/volume.py", line
139, in isLeaf
return self.getVolType() == sc.type2name(sc.LEAF_VOL)
File "/usr/lib/python2.7/site-packages/vdsm/storage/volume.py", line
135, in getVolType
self.voltype = self.getMetaParam(sc.VOLTYPE)
File "/usr/lib/python2.7/site-packages/vdsm/storage/volume.py", line
119, in getMetaParam
meta = self.getMetadata()
File "/usr/lib/python2.7/site-packages/vdsm/storage/blockVolume.py",
line 112, in getMetadata
md = VolumeMetadata.from_lines(lines)
File "/usr/lib/python2.7/site-packages/vdsm/storage/volumemetadata.py",
line 103, in from_lines
"Missing metadata key: %s: found: %s" % (e, md))
MetaDataKeyNotFoundError: Meta Data key not found error: ("Missing
metadata key: 'DOMAIN': found: {'NONE':
'######################################################################################################################################################################################################################################################################################################################################################################################################################################################################################################################'}",)
Looks like you have a volume without metadata in the chain.
This may happen in the past when deleting a volume failed, but we
cleared the volume metadata. In current 4.2, this cannot happen, since
we clear the metadata only if deleting the volume succeeded.
Can you post complete vdsm log with this error?
Once we find the volume without metadata, we can delete the LV
using lvremove. This will fix the issue.
Shani, do you remember the bug we have with this error?
this probably the same issue.
Ala, I think we need to add a tool to check and repair such chains.
Nir
--
С наилучшими пожеланиями,
Максимов Алексей
Email: Aleksey.I.Maksimov(a)Yandex.ru