----- Original Message -----
From: "Ian Levesque" <ian(a)crystal.harvard.edu>
To: "Haim Ateya" <hateya(a)redhat.com>
Cc: users(a)ovirt.org
Sent: Tuesday, May 15, 2012 10:21:04 PM
Subject: Re: [Users] Default storage pool won't activate after reboot
On May 11, 2012, at 3:09 PM, Ian Levesque wrote:
> On May 9, 2012, at 4:02 PM, Haim Ateya wrote:
>
>>>> - please access your NFS storage
>>>> - change dir to export directory - /vmstore/ovirt
>>>> - cat the following
>>>> e12a0f53-ee72-44bc-ad26-93f9b4613c6c/dom_md/metadata
>>>
>>> CLASS=Data
>>> DESCRIPTION=glusternfs
>>> IOOPTIMEOUTSEC=10
>>> LEASERETRIES=3
>>> LEASETIMESEC=60
>>> LOCKPOLICY=
>>> LOCKRENEWALINTERVALSEC=5
>>> MASTER_VERSION=1
>>> POOL_DESCRIPTION=Default
>>>
POOL_DOMAINS=82503778-db5d-4ad3-a863-b5484b2a08a1:Active,e12a0f53-ee72-44bc-ad26-93f9b4613c6c:Active
>>> POOL_SPM_ID=3
>>> POOL_SPM_LVER=5
>>> POOL_UUID=af5bcc86-898a-11e1-9632-003048c85226
>>> REMOTE_PATH=cmcd-db-vip.in.hwlab:/vmstore/ovirt
>>> ROLE=Master
>>> SDUUID=e12a0f53-ee72-44bc-ad26-93f9b4613c6c
>>> TYPE=NFS
>>> VERSION=0
>>> _SHA_CKSUM=c193c6426fd840770124c2b1db81c6138337f2ef
>>
>> the params sent by the manager connecting host to pool fits to
>> what written in metadata, hence, normally, vdsm shouldn't fail.
>> however, could please run the following:
>>
>> - mount command (on host) - if
>> 'cmcd-db-vip.in.hwlab:/vmstore/ovirt' exists, umount it.
>> - restart vdsmd service
>> - try active pool again (from manager)
>>
>> just want to make sure there is no caching issue.
>>
>> please attach logs (vdsm only) after you perform the above.
>
>
>
> I'm not even seeing the storage domain being passed to the node. I
> assume this is because the storage domain is inactive but I can't
> seem to figure out why, or what I need to do to reactivate it.
>
> The log (vdsm_log.txt) is attached, but it contains no mention of
> the storage domain. Also attached (vdsm_log_ref.txt) is a snippet
> of log that includes
> "/rhev/data-center/mnt/cmcd-db-vip.in.hwlab:_vmstore_ovirt" in the
> mount list.
Is there any further troubleshooting I can do before giving up on
this?
This is extremely odd as everything looks right, I would have blame some caching issues,
but you mentioned
that vdsm service was restarted.
Saggi\Edu\Federico - any insight on this case ? why would connectStoragePool fails on
cannot find master domain if command
sent by manager uses right params as mentioned in meta-data ? connectStorageServer
succeeds as well (although I see no mount command - guess it due to OOP).
Ian - you can always use the "Re-initialize data-center" flow which basically
means you loss all your current data (iirc) and new master domain is created.
let's wait for the guys to give their insight, meantime I would try reboot the host
just to make sure there is no stale NFS entry there.
Thanks,
Ian