[ovirt-users] _initialize_sanlock cannot get lock, host already holds lock on a different host id

Robert Story rstory at tislabs.com
Thu Nov 12 18:12:49 UTC 2015


On Thu, 12 Nov 2015 12:54:49 -0500 Robert wrote:
RS> On Thu, 12 Nov 2015 15:22:18 +0100 Sandro wrote:
RS> SB> > I'm running oVirt 3.5.x with a hosted engine. On 3 of my 5 nodes,
RS> SB> > ovirt-ha-agent won't start, complaining that
RS> SB> > "(_initialize_sanlock) cannot get lock on host id 5: host already
RS> SB> > holds lock on a different host id."
RS> SB> >
RS> SB> >
RS> SB> It should correctly refuse to start the vm since the lock is already
RS> SB> taken, not sure if the message log is just confusing or a real
RS> SB> issue.
RS> 
RS> Just to clarify, this isn't about a vm. The engine VM is up and I'm not
RS> having issues with any other vms. The problem is with the
RS> ovirt-ha-agent.

Some additional info. I ran 'sanlock client status -D' on 1 working and 1
non-working host.

host 3 (working):
s hosted-engine:3:/var/run/vdsm/storage/2daba0ab-2b3d-4026-bcfc-1cd071c30038/04b08c8e-657f-4bac-9ddf-c9c57373409c/2d7f5020-42c1-442d-8237-fba9d6787080:0
    list=spaces
    space_id=4
    io_timeout=10
    host_generation=5
    renew_fail=0
    space_dead=0
    killing_pids=0
    used_retries=0
    external_used=0
    used_by_orphans=0
    corrupt_result=0
    acquire_last_result=1
    renewal_last_result=1
    acquire_last_attempt=2178388
    acquire_last_success=2178528
    renewal_last_attempt=3523708
    renewal_last_success=3523708

host 5 (not working):
s hosted-engine:5:/rhev/data-center/mnt/ovirt-nfs.netsec\:_ovirt_hosted-engine/2daba0ab-2b3d-4026-bcfc-1cd071c30038/images/04b08c8e-657f-4bac-9ddf-c9c57373409c/2d7f5020-42c1-442d-8237-fba9d6787080:0
    list=spaces
    space_id=2
    io_timeout=10
    host_generation=17
    renew_fail=0
    space_dead=0
    killing_pids=0
    used_retries=0
    external_used=0
    used_by_orphans=0
    corrupt_result=0
    acquire_last_result=1
    renewal_last_result=1
    acquire_last_attempt=101
    acquire_last_success=241
    renewal_last_attempt=3532404
    renewal_last_success=3532404

And running 'sanlock client host_status -s hosted-engine -D' (on either 3
or 5), info for hosts 3 and 5 is:

3 timestamp 3523933
    last_check=3523954
    last_live=3523954
    last_req=0
    owner_id=3
    owner_generation=5
    timestamp=3523933
    io_timeout=10
    owner_name=53d2cee3-fdd8-4c4c-8265-83328bf729af.eclipse.ne
5 timestamp 3532732
    last_check=3523954
    last_live=3523954
    last_req=0
    owner_id=5
    owner_generation=17
    timestamp=3532732
    io_timeout=10
    owner_name=2c1ec955-4802-4f89-a824-d8a7470c2c9f.apollo.net

Robert

-- 
Senior Software Engineer @ Parsons
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 819 bytes
Desc: OpenPGP digital signature
URL: <http://lists.ovirt.org/pipermail/users/attachments/20151112/2af6a009/attachment-0001.sig>


More information about the Users mailing list