Re: [ClusterLabs] Antw: Re: Disabled resource is hard logging

[Oscar Segarra]

Hi,

After applying last changes... it looks not appear de huge amount of errors
related to vdicone01 vm.

I apologice for my maybe simple question, but can you explain which is the
difference between the following commands:

pcs resource op remove vm-vdicone01 monitor role=Stopped
pcs resource op remove vm-vdicone01 stop interval=0s timeout=90

After executing both commands I have experienced that sometimes (not
always) in virt-manager I can see the vdicone01 started on hypervisor1 and
stopped in hypervisor2. I can delete it from hypervisor2 (not deleting
storage) but it appears. This behaviour can be caused by that commands?

Thanks in advance!

2017-02-17 8:33 GMT+01:00 Ulrich Windl :

> >>> Oscar Segarra  schrieb am 16.02.2017 um
> 13:55 in
> Nachricht
> :
> > Hi Klaus,
> >
> > Thanks a lot, I will try to delete the stop monitor.
> >
> > Nevertheless, I have 6 domains configured exactly the same... Is there
> any
> > reason why just this domain has this behaviour ?
>
> Some years ago I was playing with NPIV, and it worked perfectly for one and
> for several VMs. However when multiple VMs were started or stopped at the
> same
> time (this NPIV added/removed), I had "interesting" failures due to
> concurrency, even a kernel lockup (which is fixed meanwhile). So most
> likely
> "something is not correct".
> I know it doesn't help you the way you would like, but it's how life is.
>
> Regards,
> Ulrich
>
> >
> > Thanks a lot.
> >
> > 2017-02-16 11:12 GMT+01:00 Klaus Wenninger :
> >
> >> On 02/16/2017 11:02 AM, Oscar Segarra wrote:
> >> > Hi Kaluss
> >> >
> >> > Which is your proposal to fix this behavior?
> >>
> >> First you can try to remove the monitor op for role=stopped.
> >> Then the startup-probing will probably still fail but for that
> >> the behaviour is different.
> >> The startup-probing can be disabled globally via cluster-property
> >> enable-startup-probes that is defaulting to true.
> >> But be aware that the cluster then wouldn't be able to react
> >> properly if services are already up when pacemaker is starting.
> >> It should be possible to disable the probing on a per resource
> >> or node basis as well iirc. But I can't tell you out of my mind
> >> how that worked - there was a discussion a few weeks ago
> >> on the list iirc.
> >>
> >> Regards,
> >> Klaus
> >>
> >> >
> >> > Thanks a lot!
> >> >
> >> >
> >> > El 16 feb. 2017 10:57 a. m., "Klaus Wenninger"  >> > > escribió:
> >> >
> >> > On 02/16/2017 09:05 AM, Oscar Segarra wrote:
> >> > > Hi,
> >> > >
> >> > > In my environment I have deployed 5 VirtualDomains as one can
> >> > see below:
> >> > > [root@vdicnode01 ~]# pcs status
> >> > > Cluster name: vdic-cluster
> >> > > Stack: corosync
> >> > > Current DC: vdicnode01-priv (version 1.1.15-11.el7_3.2-e174ec8)
> -
> >> > > partition with quorum
> >> > > Last updated: Thu Feb 16 09:02:53 2017  Last change: Thu
> >> Feb
> >> > > 16 08:20:53 2017 by root via crm_attribute on vdicnode02-priv
> >> > >
> >> > > 2 nodes and 14 resources configured: 5 resources DISABLED and 0
> >> > > BLOCKED from being started due to failures
> >> > >
> >> > > Online: [ vdicnode01-priv vdicnode02-priv ]
> >> > >
> >> > > Full list of resources:
> >> > >
> >> > >  nfs-vdic-mgmt-vm-vip   (ocf::heartbeat:IPaddr):Started
> >> > > vdicnode01-priv
> >> > >  Clone Set: nfs_setup-clone [nfs_setup]
> >> > >  Started: [ vdicnode01-priv vdicnode02-priv ]
> >> > >  Clone Set: nfs-mon-clone [nfs-mon]
> >> > >  Started: [ vdicnode01-priv vdicnode02-priv ]
> >> > >  Clone Set: nfs-grace-clone [nfs-grace]
> >> > >  Started: [ vdicnode01-priv vdicnode02-priv ]
> >> > >  vm-vdicone01   (ocf::heartbeat:VirtualDomain): FAILED
> (disabled)[
> >> > > vdicnode02-priv vdicnode01-priv ]
> >> > >  vm-vdicsunstone01  (ocf::heartbeat:VirtualDomain): FAILED
> >> > > vdicnode01-priv (disabled)
> >> > >  vm-vdicdb01(ocf::heartbeat:VirtualDomain): FAILED
> (disabled)[
> >> > > vdicnode02-priv vdicnode01-priv ]
> >> > >  vm-vdicudsserver   (ocf::heartbeat:VirtualDomain): FAILED
> >> > > (disabled)[ vdicnode02-priv vdicnode01-priv ]
> >> > >  vm-vdicudstuneler  (ocf::heartbeat:VirtualDomain): FAILED
> >> > > vdicnode01-priv (disabled)
> >> > >  Clone Set: nfs-vdic-images-vip-clone [nfs-vdic-images-vip]
> >> > >  Stopped: [ vdicnode01-priv vdicnode02-priv ]
> >> > >
> >> > > Failed Actions:
> >> > > * vm-vdicone01_monitor_2 on vdicnode02-priv 'not installed'
> >> (5):
> >> > > call=2322, status=complete, exitreason='Configuration file
> >> > > /mnt/nfs-vdic-mgmt-vm/vdicone01.xml does not exist or is not
> >> > 

[ClusterLabs] Antw: Re: Disabled resource is hard logging

[Ulrich Windl]

>>> Oscar Segarra  schrieb am 16.02.2017 um 13:55 in
Nachricht
:
> Hi Klaus,
> 
> Thanks a lot, I will try to delete the stop monitor.
> 
> Nevertheless, I have 6 domains configured exactly the same... Is there any
> reason why just this domain has this behaviour ?

Some years ago I was playing with NPIV, and it worked perfectly for one and
for several VMs. However when multiple VMs were started or stopped at the same
time (this NPIV added/removed), I had "interesting" failures due to
concurrency, even a kernel lockup (which is fixed meanwhile). So most likely
"something is not correct".
I know it doesn't help you the way you would like, but it's how life is.

Regards,
Ulrich

> 
> Thanks a lot.
> 
> 2017-02-16 11:12 GMT+01:00 Klaus Wenninger :
> 
>> On 02/16/2017 11:02 AM, Oscar Segarra wrote:
>> > Hi Kaluss
>> >
>> > Which is your proposal to fix this behavior?
>>
>> First you can try to remove the monitor op for role=stopped.
>> Then the startup-probing will probably still fail but for that
>> the behaviour is different.
>> The startup-probing can be disabled globally via cluster-property
>> enable-startup-probes that is defaulting to true.
>> But be aware that the cluster then wouldn't be able to react
>> properly if services are already up when pacemaker is starting.
>> It should be possible to disable the probing on a per resource
>> or node basis as well iirc. But I can't tell you out of my mind
>> how that worked - there was a discussion a few weeks ago
>> on the list iirc.
>>
>> Regards,
>> Klaus
>>
>> >
>> > Thanks a lot!
>> >
>> >
>> > El 16 feb. 2017 10:57 a. m., "Klaus Wenninger" > > > escribió:
>> >
>> > On 02/16/2017 09:05 AM, Oscar Segarra wrote:
>> > > Hi,
>> > >
>> > > In my environment I have deployed 5 VirtualDomains as one can
>> > see below:
>> > > [root@vdicnode01 ~]# pcs status
>> > > Cluster name: vdic-cluster
>> > > Stack: corosync
>> > > Current DC: vdicnode01-priv (version 1.1.15-11.el7_3.2-e174ec8) -
>> > > partition with quorum
>> > > Last updated: Thu Feb 16 09:02:53 2017  Last change: Thu
>> Feb
>> > > 16 08:20:53 2017 by root via crm_attribute on vdicnode02-priv
>> > >
>> > > 2 nodes and 14 resources configured: 5 resources DISABLED and 0
>> > > BLOCKED from being started due to failures
>> > >
>> > > Online: [ vdicnode01-priv vdicnode02-priv ]
>> > >
>> > > Full list of resources:
>> > >
>> > >  nfs-vdic-mgmt-vm-vip   (ocf::heartbeat:IPaddr):Started
>> > > vdicnode01-priv
>> > >  Clone Set: nfs_setup-clone [nfs_setup]
>> > >  Started: [ vdicnode01-priv vdicnode02-priv ]
>> > >  Clone Set: nfs-mon-clone [nfs-mon]
>> > >  Started: [ vdicnode01-priv vdicnode02-priv ]
>> > >  Clone Set: nfs-grace-clone [nfs-grace]
>> > >  Started: [ vdicnode01-priv vdicnode02-priv ]
>> > >  vm-vdicone01   (ocf::heartbeat:VirtualDomain): FAILED (disabled)[
>> > > vdicnode02-priv vdicnode01-priv ]
>> > >  vm-vdicsunstone01  (ocf::heartbeat:VirtualDomain): FAILED
>> > > vdicnode01-priv (disabled)
>> > >  vm-vdicdb01(ocf::heartbeat:VirtualDomain): FAILED (disabled)[
>> > > vdicnode02-priv vdicnode01-priv ]
>> > >  vm-vdicudsserver   (ocf::heartbeat:VirtualDomain): FAILED
>> > > (disabled)[ vdicnode02-priv vdicnode01-priv ]
>> > >  vm-vdicudstuneler  (ocf::heartbeat:VirtualDomain): FAILED
>> > > vdicnode01-priv (disabled)
>> > >  Clone Set: nfs-vdic-images-vip-clone [nfs-vdic-images-vip]
>> > >  Stopped: [ vdicnode01-priv vdicnode02-priv ]
>> > >
>> > > Failed Actions:
>> > > * vm-vdicone01_monitor_2 on vdicnode02-priv 'not installed'
>> (5):
>> > > call=2322, status=complete, exitreason='Configuration file
>> > > /mnt/nfs-vdic-mgmt-vm/vdicone01.xml does not exist or is not
>> > readable.',
>> > > last-rc-change='Thu Feb 16 09:02:07 2017', queued=0ms,
>> exec=21ms
>> > > * vm-vdicsunstone01_monitor_2 on vdicnode02-priv 'not
>> installed'
>> > > (5): call=2310, status=complete, exitreason='Configuration file
>> > > /mnt/nfs-vdic-mgmt-vm/vdicsunstone01.xml does not exist or is not
>> > > readable.',
>> > > last-rc-change='Thu Feb 16 09:02:07 2017', queued=0ms,
>> exec=37ms
>> > > * vm-vdicdb01_monitor_2 on vdicnode02-priv 'not installed'
(5):
>> > > call=2320, status=complete, exitreason='Configuration file
>> > > /mnt/nfs-vdic-mgmt-vm/vdicdb01.xml does not exist or is not
>> > readable.',
>> > > last-rc-change='Thu Feb 16 09:02:07 2017', queued=0ms,
>> exec=35ms
>> > > * vm-vdicudsserver_monitor_2 on vdicnode02-priv 'not
installed'
>> > > (5): call=2321, status=complete, exitreason='Configuration file
>> > >