[ovirt-users] Importing existing (dirty) storage domains

Fri Feb 17 02:15:46 UTC 2017

Hi Nir,

On 16 Feb 2017 22:41, "Nir Soffer" <nsoffer at redhat.com> wrote:

On Fri, Feb 17, 2017 at 3:16 AM, Doug Ingham <dougti at gmail.com> wrote:
> Well that didn't go so well. I deleted both dom_md/ids & dom_md/leases in
> the cloned volume, and I still can't import the storage domain.

You cannot delete files from dom_md, this will invalidate the storage
domain and you
will not be able to use it without restoring these files.

The leases file on file storage domain is unused, so creating empty
file is enough.

The ids file must be created and initialized using sanlock, please you
should find
instructions how to do it in the archives.

> The snapshot was also taken some 4 hours before the attempted import, so
I'm
> surprised the locks haven't expired by themselves...

Leases do not expire if vdsm is connected to storage, and sanlock can access
the storage.

I'm not sure what do you mean by volume snapshots.

A snapshot is like a save-point, the state of a storage volume from a
specific point in time.

In this case, it means I have created a copy/clone of my active data
volume. It's a completely new, separate volume, and is not attached to any
running services.

I'm using this copy/clone to test the import process, before doing it with
my live volume. If I "break" something in the cloned volume, no worries, I
can just delete it and recreate it from the snapshot.

Hope that clears things up a bit!

To import a storage domain, you should first make that no other setup is
using
the storage domain.

The best way to do it is to detach the storage domain from the other setup.

If you are using hosted engine, you must also put the hosted engine agent in
global maintenance mode.

If your engine is broken, the best way to disconnect from storage is to
reboot
the hosts.

So this is the issue. My current tests emulate exactly this, however I'm
still not able to import the domain into the new Engine. When I try to do
so, I get the resulting logs I copied in my earlier email.

Doug

Nir

>
>
> 2017-02-16 21:58:24,630-03 INFO
> [org.ovirt.engine.core.bll.storage.connection.
AddStorageServerConnectionCommand]
> (default task-45) [d59bc8c0-3c53-4a34-9d7c-8c982ee14e14] Lock Acquired to
> object
> 'EngineLock:{exclusiveLocks='[localhost:data-teste2=<STORAGE_CONNECTION,
> ACTION_TYPE_FAILED_OBJECT_LOCKED>]', sharedLocks='null'}'
> 2017-02-16 21:58:24,645-03 INFO
> [org.ovirt.engine.core.vdsbroker.vdsbroker.ConnectStorageServerVDSCommand]
> (default task-45) [d59bc8c0-3c53-4a34-9d7c-8c982ee14e14] START,
> ConnectStorageServerVDSCommand(HostName = v5.dc0.example.com,
> StorageServerConnectionManagementVDSParameters:{runAsync='true',
> hostId='1a3f10f2-e4ce-44b9-9495-06e445cfa0b0',
> storagePoolId='00000000-0000-0000-0000-000000000000',
> storageType='GLUSTERFS',
> connectionList='[StorageServerConnections:{id='null',
> connection='localhost:data-teste2', iqn='null', vfsType='glusterfs',
> mountOptions='null', nfsVersion='null', nfsRetrans='null',
nfsTimeo='null',
> iface='null', netIfaceName='null'}]'}), log id: 726df65e
> 2017-02-16 21:58:26,046-03 INFO
> [org.ovirt.engine.core.bll.storage.connection.
AddStorageServerConnectionCommand]
> (default task-45) [d59bc8c0-3c53-4a34-9d7c-8c982ee14e14] Lock freed to
> object 'EngineLock:{exclusiveLocks='[localhost:data
> teste2=<STORAGE_CONNECTION, ACTION_TYPE_FAILED_OBJECT_LOCKED>]',
> sharedLocks='null'}'
> 2017-02-16 21:58:26,206-03 INFO
> [org.ovirt.engine.core.vdsbroker.vdsbroker.HSMGetStorageDomainsListVDSCom
mand]
> (default task-52) [85548427-713f-4ffb-a385-a97a7ee4109d] START,
> HSMGetStorageDomainsListVDSCommand(HostName = v5.dc0.example.com,
> HSMGetStorageDomainsListVDSCommandParameters:{runAsync='true',
> hostId='1a3f10f2-e4ce-44b9-9495-06e445cfa0b0',
> storagePoolId='00000000-0000-0000-0000-000000000000', storageType='null',
> storageDomainType='Data', path='localhost:data-teste2'}), log id: 79f6cc88
> 2017-02-16 21:58:27,899-03 INFO
> [org.ovirt.engine.core.vdsbroker.vdsbroker.HSMGetStorageDomainsListVDSCom
mand]
> (default task-50) [38e87311-a7a5-49a8-bf18-857dd969cd5f] START,
> HSMGetStorageDomainsListVDSCommand(HostName = v5.dc0.example.com,
> HSMGetStorageDomainsListVDSCommandParameters:{runAsync='true',
> hostId='1a3f10f2-e4ce-44b9-9495-06e445cfa0b0',
> storagePoolId='00000000-0000-0000-0000-000000000000', storageType='null',
> storageDomainType='Data', path='localhost:data-teste2'}), log id: 7280d13
> 2017-02-16 21:58:29,156-03 INFO
> [org.ovirt.engine.core.bll.storage.connection.
RemoveStorageServerConnectionCommand]
> (default task-56) [1b3826e4-4890-43d4-8854-16f3c573a31f] Lock Acquired to
> object
> 'EngineLock:{exclusiveLocks='[localhost:data-teste2=<STORAGE_CONNECTION,
> ACTION_TYPE_FAILED_OBJECT_LOCKED>,
> 5e5f6610-c759-448b-a53d-9a456f513681=<STORAGE_CONNECTION,
> ACTION_TYPE_FAILED_OBJECT_LOCKED>]', sharedLocks='null'}'
> 2017-02-16 21:58:29,168-03 INFO
> [org.ovirt.engine.core.vdsbroker.vdsbroker.DisconnectStorageServerVDSComm
and]
> (default task-57) [5e4b20cf-60d2-4ae9-951b-c2693603aa6f] START,
> DisconnectStorageServerVDSCommand(HostName = v5.dc0.example.com,
> StorageServerConnectionManagementVDSParameters:{runAsync='true',
> hostId='1a3f10f2-e4ce-44b9-9495-06e445cfa0b0',
> storagePoolId='00000000-0000-0000-0000-000000000000',
> storageType='GLUSTERFS',
> connectionList='[StorageServerConnections:{id='5e5f6610-c759-448b-a53d-
9a456f513681',
> connection='localhost:data-teste2', iqn='null', vfsType='glusterfs',
> mountOptions='null', nfsVersion='null', nfsRetrans='null',
nfsTimeo='null',
> iface='null', netIfaceName='null'}]'}), log id: 6042b108
> 2017-02-16 21:58:29,193-03 INFO
> [org.ovirt.engine.core.vdsbroker.vdsbroker.DisconnectStorageServerVDSComm
and]
> (default task-56) [1b3826e4-4890-43d4-8854-16f3c573a31f] START,
> DisconnectStorageServerVDSCommand(HostName = v5.dc0.example.com,
> StorageServerConnectionManagementVDSParameters:{runAsync='true',
> hostId='1a3f10f2-e4ce-44b9-9495-06e445cfa0b0',
> storagePoolId='00000000-0000-0000-0000-000000000000',
> storageType='GLUSTERFS',
> connectionList='[StorageServerConnections:{id='5e5f6610-c759-448b-a53d-
9a456f513681',
> connection='localhost:data-teste2', iqn='null', vfsType='glusterfs',
> mountOptions='null', nfsVersion='null', nfsRetrans='null',
nfsTimeo='null',
> iface='null', netIfaceName='null'}]'}), log id: 4e9421cf
> 2017-02-16 21:58:31,398-03 INFO
> [org.ovirt.engine.core.bll.storage.connection.
RemoveStorageServerConnectionCommand]
> (default task-56) [1b3826e4-4890-43d4-8854-16f3c573a31f] Lock freed to
> object
> 'EngineLock:{exclusiveLocks='[localhost:data-teste2=<STORAGE_CONNECTION,
> ACTION_TYPE_FAILED_OBJECT_LOCKED>,
> 5e5f6610-c759-448b-a53d-9a456f513681=<STORAGE_CONNECTION,
> ACTION_TYPE_FAILED_OBJECT_LOCKED>]', sharedLocks='null'}'
>
> Again, many thanks!
>  Doug
>
> On 16 February 2017 at 18:53, Doug Ingham <dougti at gmail.com> wrote:
>>
>> Hi Nir,
>>
>> On 16 February 2017 at 13:55, Nir Soffer <nsoffer at redhat.com> wrote:
>>>
>>> On Mon, Feb 13, 2017 at 3:35 PM, Doug Ingham <dougti at gmail.com> wrote:
>>> > Hi Sahina,
>>> >
>>> > On 13 February 2017 at 05:45, Sahina Bose <sabose at redhat.com> wrote:
>>> >>
>>> >> Any errors in the gluster mount logs for this gluster volume?
>>> >>
>>> >> How about "gluster vol heal <volname> info" - does it list any
entries
>>> >> to
>>> >> heal?
>>> >
>>> >
>>> > After more investigating, I found out that there is a sanlock daemon
>>> > that
>>> > runs with VDSM, independently of the HE, so I'd basically have to
bring
>>> > the
>>> > volume down & wait for the leases to expire/delete them* before I can
>>> > import
>>> > the domain.
>>> >
>>> > *I understand removing /dom_md/leases/ should do the job?
>>>
>>> No, the issue is probably dom_md/ids accessed by sanlock, but removing
>>> files
>>> accessed by sanlock will not help, an open file will remain open until
>>> sanlock
>>> close the file.
>>
>>
>> I'm testing this with volume snapshots at the moment, so there are no
>> processes accessing the new volume.
>>
>>
>>> Did you try to reboot the host before installing it again? If you did
and
>>> you
>>> still have these issues, you probably need to remove the previous
>>> installation
>>> properly before installing again.
>>>
>>> Adding Simone to help with uninstalling and reinstalling hosted engine.
>>
>>
>> The Hosted-Engine database had been corrupted and the restore wasn't
>> running correctly, so I installed a new engine on a new server - no
restores
>> or old data. The aim is to import the old storage domain into the new
Engine
>> & then import the VMs into the new storage domain.
>> My only problem with this is that there appear to be some file based
>> leases somewhere that, unless I manage to locate & delete them, force me
to
>> wait for the leases to timeout before I can import the old storage
domain.
>> To minimise downtime, I'm trying to avoid having to wait for the leases
to
>> timeout.
>>
>> Regards,
>>  Doug
>>
>>>
>>>
>>> Nir
>>>
>>> >
>>> >
>>> >>
>>> >>
>>> >> On Thu, Feb 9, 2017 at 11:57 PM, Doug Ingham <dougti at gmail.com>
wrote:
>>> >>>
>>> >>> Some interesting output from the vdsm log...
>>> >>>
>>> >>>
>>> >>> 2017-02-09 15:16:24,051 INFO  (jsonrpc/1) [storage.StorageDomain]
>>> >>> Resource namespace 01_img_60455567-ad30-42e3-a9df-62fe86c7fd25
>>> >>> already
>>> >>> registered (sd:731)
>>> >>> 2017-02-09 15:16:24,051 INFO  (jsonrpc/1) [storage.StorageDomain]
>>> >>> Resource namespace 02_vol_60455567-ad30-42e3-a9df-62fe86c7fd25
>>> >>> already
>>> >>> registered (sd:740)
>>> >>> 2017-02-09 15:16:24,052 INFO  (jsonrpc/1) [storage.SANLock]
Acquiring
>>> >>> Lease(name='SDM',
>>> >>>
>>> >>> path=u'/rhev/data-center/mnt/glusterSD/localhost:data2/
60455567-ad30-42e3-a9df-6
>>> >>> 2fe86c7fd25/dom_md/leases', offset=1048576) for host id 1
>>> >>> (clusterlock:343)
>>> >>> 2017-02-09 15:16:24,057 INFO  (jsonrpc/1) [storage.SANLock]
Releasing
>>> >>> host id for domain 60455567-ad30-42e3-a9df-62fe86c7fd25 (id: 1)
>>> >>> (clusterlock:305)
>>> >>> 2017-02-09 15:16:25,149 INFO  (jsonrpc/3) [jsonrpc.JsonRpcServer]
RPC
>>> >>> call GlusterHost.list succeeded in 0.17 seconds (__init__:515)
>>> >>> 2017-02-09 15:16:25,264 INFO  (Reactor thread)
>>> >>> [ProtocolDetector.AcceptorImpl] Accepted connection from
>>> >>> ::ffff:127.0.0.1:55060 (protocoldetector:72)
>>> >>> 2017-02-09 15:16:25,270 INFO  (Reactor thread)
>>> >>> [ProtocolDetector.Detector] Detected protocol stomp from
>>> >>> ::ffff:127.0.0.1:55060 (protocoldetector:127)
>>> >>> 2017-02-09 15:16:25,271 INFO  (Reactor thread) [Broker.StompAdapter]
>>> >>> Processing CONNECT request (stompreactor:102)
>>> >>> 2017-02-09 15:16:25,271 INFO  (JsonRpc (StompReactor))
>>> >>> [Broker.StompAdapter] Subscribe command received (stompreactor:129)
>>> >>> 2017-02-09 15:16:25,416 INFO  (jsonrpc/5) [jsonrpc.JsonRpcServer]
RPC
>>> >>> call Host.getHardwareInfo succeeded in 0.01 seconds (__init__:515)
>>> >>> 2017-02-09 15:16:25,419 INFO  (jsonrpc/6) [dispatcher] Run and
>>> >>> protect:
>>> >>> repoStats(options=None) (logUtils:49)
>>> >>> 2017-02-09 15:16:25,419 INFO  (jsonrpc/6) [dispatcher] Run and
>>> >>> protect:
>>> >>> repoStats, Return response: {u'e8d04da7-ad3d-4227-a45d-
b5a29b2f43e5':
>>> >>> {'code': 0, 'actual': True
>>> >>> , 'version': 4, 'acquired': True, 'delay': '0.000854128',
>>> >>> 'lastCheck':
>>> >>> '5.1', 'valid': True}, u'a77b8821-ff19-4d17-a3ce-a6c3a69436d5':
>>> >>> {'code': 0,
>>> >>> 'actual': True, 'vers
>>> >>> ion': 4, 'acquired': True, 'delay': '0.000966556', 'lastCheck':
>>> >>> '2.6',
>>> >>> 'valid': True}} (logUtils:52)
>>> >>> 2017-02-09 15:16:25,447 INFO  (jsonrpc/6) [jsonrpc.JsonRpcServer]
RPC
>>> >>> call Host.getStats succeeded in 0.03 seconds (__init__:515)
>>> >>> 2017-02-09 15:16:25,450 ERROR (JsonRpc (StompReactor))
>>> >>> [vds.dispatcher]
>>> >>> SSL error receiving from <yajsonrpc.betterAsyncore.Dispatcher
>>> >>> connected
>>> >>> ('::ffff:127.0.0.1', 55060, 0, 0) at 0x7f69c0043cf8>: unexpected eof
>>> >>> (betterAsyncore:113)
>>> >>> 2017-02-09 15:16:25,812 INFO  (jsonrpc/7) [jsonrpc.JsonRpcServer]
RPC
>>> >>> call GlusterVolume.list succeeded in 0.10 seconds (__init__:515)
>>> >>> 2017-02-09 15:16:25,940 INFO  (Reactor thread)
>>> >>> [ProtocolDetector.AcceptorImpl] Accepted connection from
>>> >>> ::ffff:127.0.0.1:55062 (protocoldetector:72)
>>> >>> 2017-02-09 15:16:25,946 INFO  (Reactor thread)
>>> >>> [ProtocolDetector.Detector] Detected protocol stomp from
>>> >>> ::ffff:127.0.0.1:55062 (protocoldetector:127)
>>> >>> 2017-02-09 15:16:25,947 INFO  (Reactor thread) [Broker.StompAdapter]
>>> >>> Processing CONNECT request (stompreactor:102)
>>> >>> 2017-02-09 15:16:25,947 INFO  (JsonRpc (StompReactor))
>>> >>> [Broker.StompAdapter] Subscribe command received (stompreactor:129)
>>> >>> 2017-02-09 15:16:26,058 ERROR (jsonrpc/1) [storage.TaskManager.Task]
>>> >>> (Task='02cad901-5fe8-4f2d-895b-14184f67feab') Unexpected error
>>> >>> (task:870)
>>> >>> Traceback (most recent call last):
>>> >>>   File "/usr/share/vdsm/storage/task.py", line 877, in _run
>>> >>>     return fn(*args, **kargs)
>>> >>>   File "/usr/lib/python2.7/site-packages/vdsm/logUtils.py", line 50,
>>> >>> in
>>> >>> wrapper
>>> >>>     res = f(*args, **kwargs)
>>> >>>   File "/usr/share/vdsm/storage/hsm.py", line 812, in
>>> >>> forcedDetachStorageDomain
>>> >>>     self._deatchStorageDomainFromOldPools(sdUUID)
>>> >>>   File "/usr/share/vdsm/storage/hsm.py", line 790, in
>>> >>> _deatchStorageDomainFromOldPools
>>> >>>     dom.acquireClusterLock(host_id)
>>> >>>   File "/usr/share/vdsm/storage/sd.py", line 810, in
>>> >>> acquireClusterLock
>>> >>>     self._manifest.acquireDomainLock(hostID)
>>> >>>   File "/usr/share/vdsm/storage/sd.py", line 499, in
>>> >>> acquireDomainLock
>>> >>>     self._domainLock.acquire(hostID, self.getDomainLease())
>>> >>>   File
>>> >>> "/usr/lib/python2.7/site-packages/vdsm/storage/clusterlock.py",
>>> >>> line 362, in acquire
>>> >>>     "Cannot acquire %s" % (lease,), str(e))
>>> >>> AcquireLockFailure: Cannot obtain lock:
>>> >>> u"id=60455567-ad30-42e3-a9df-62fe86c7fd25, rc=5, out=Cannot acquire
>>> >>> Lease(name='SDM',
>>> >>>
>>> >>> path=u'/rhev/data-center/mnt/glusterSD/localhost:data2/
60455567-ad30-42e3-a9df-62fe86c7fd25/dom_md/leases',
>>> >>> offset=1048576), err=(5, 'Sanlock resource not acquired',
>>> >>> 'Input/output
>>> >>> error')"
>>> >>> 2017-02-09 15:16:26,058 INFO  (jsonrpc/1) [storage.TaskManager.Task]
>>> >>> (Task='02cad901-5fe8-4f2d-895b-14184f67feab') aborting: Task is
>>> >>> aborted:
>>> >>> 'Cannot obtain lock' - code 651 (task:1175)
>>> >>> 2017-02-09 15:16:26,059 ERROR (jsonrpc/1) [storage.Dispatcher]
>>> >>> {'status':
>>> >>> {'message': 'Cannot obtain lock:
>>> >>> u"id=60455567-ad30-42e3-a9df-62fe86c7fd25,
>>> >>> rc=5, out=Cannot acquire Lease(name=\'SDM\',
>>> >>>
>>> >>> path=u\'/rhev/data-center/mnt/glusterSD/localhost:data2/
60455567-ad30-42e3-a9df-62fe86c7fd25/dom_md/leases\',
>>> >>> offset=1048576), err=(5, \'Sanlock resource not acquired\',
>>> >>> \'Input/output
>>> >>> error\')"', 'code': 651}} (dispatcher:77)
>>> >>> 2017-02-09 15:16:26,059 INFO  (jsonrpc/1) [jsonrpc.JsonRpcServer]
RPC
>>> >>> call StorageDomain.detach failed (error 651) in 23.04 seconds
>>> >>> (__init__:515)
>>> >>>
>>> >>> --
>>> >>> Doug
>>> >>>
>>> >>> _______________________________________________
>>> >>> Users mailing list
>>> >>> Users at ovirt.org
>>> >>> http://lists.ovirt.org/mailman/listinfo/users
>>> >>>
>>> >>
>>> >
>>> >
>>> >
>>> > --
>>> > Doug
>>> >
>>> > _______________________________________________
>>> > Users mailing list
>>> > Users at ovirt.org
>>> > http://lists.ovirt.org/mailman/listinfo/users
>>> >
>>
>>
>>
>>
>> --
>> Doug
>
>
>
>
> --
> Doug
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ovirt.org/pipermail/users/attachments/20170216/e05d34ec/attachment-0001.html>