Hi,
On 11/27/2015 01:58 PM, ml wrote:> Dear All,
>
> I am trying to get a nfs-ganesha ha cluster running, with 3, CentOS
> Linux release 7.1.1503 nodes. I use the package glusterfs-ganesha-3.7.6
> -1.el7.x86_64 to get the HA scripts. So far it works fine when i stop
> the nfs-ganesha service on one of the node it moves the virtual ip to
> one of the other node, altai-dead_ip-1 resource is created properly:
>
>
> root at rnas2 ~# pcs status
> Cluster name: ganesha-cluster-dmath
> Last updated: Thu Nov 26 10:41:07 2015 Last
> change: Thu Nov 26 10:40:06 2015 by root via cibadmin on altai
> Stack: corosync
> Current DC: rnas2 (version 1.1.13-a14efad) - partition with
> quorum
> 3 nodes and 13 resources configured
>
> Online: [ altai kaukasus rnas2 ]
>
> Full list of resources:
>
> Clone Set: nfs-mon-clone [nfs-mon]
> Started: [ altai kaukasus rnas2 ]
> Clone Set: nfs-grace-clone [nfs-grace]
> Started: [ altai kaukasus rnas2 ]
> kaukasus-cluster_ip-1 (ocf::heartbeat:IPaddr): S
> tarted kaukasus
> kaukasus-trigger_ip-1 (ocf::heartbeat:Dummy): St
> arted kaukasus
> altai-cluster_ip-1 (ocf::heartbeat:IPaddr): Star
> ted kaukasus
> altai-trigger_ip-1 (ocf::heartbeat:Dummy): Start
> ed kaukasus
> rnas2-cluster_ip-1 (ocf::heartbeat:IPaddr): Star
> ted rnas2
> rnas2-trigger_ip-1 (ocf::heartbeat:Dummy): Start
> ed rnas2
> altai-dead_ip-1 (ocf::heartbeat:Dummy): Started
> altai
>
> PCSD Status:
> kaukasus: Online
> altai: Online
> rnas2: Online
>
> Daemon Status:
> corosync: active/enabled
> pacemaker: active/enabled
> pcsd: active/enabled
>
>
> But when i just disconnect the network on one of the node, in this case
> altai (or poweroff),
>
>
> root at altai ~# ifdown bond0
>
>
> it takes down the whole cluster. I found the following message in the
> logs:
>
>
> Nov 26 10:45:05 rnas2 crmd[17255]: error: Operation nfs
> -grace_start_0: Timed Out (node=rnas2, call=85, timeout=40000ms)
>
>
> I wonder if i just misconfigured something or if this is not supported
> yet?
>
Since its a 3-node cluster, quorum shall be enabled. When any of those
machine/its IP is down, quorum shall be lost resulting in pacemaker
shutting down entire cluster. If possible could you check the same
scenario with 4-node setup?
Thanks,
Soumya>
> below the log during the take down:
>
> Nov 26 10:44:24 rnas2 corosync[8848]: [TOTEM ] A new membership
> (129.132.145.5:1048) was formed. Members left: 2
> Nov 26 10:44:24 rnas2 attrd[17253]: notice:
> crm_update_peer_proc: Node altai[2] - state is now lost (was member)
> Nov 26 10:44:24 rnas2 attrd[17253]: notice: Removing all altai
> attributes for attrd_peer_change_cb
> Nov 26 10:44:25 rnas2 corosync[8848]: [QUORUM] Members[2]: 1 3
> Nov 26 10:44:25 rnas2 corosync[8848]: [MAIN ] Completed
> service synchronization, ready to provide service.
> Nov 26 10:44:25 rnas2 cib[17250]: notice: crm_update_peer_proc:
> Node altai[2] - state is now lost (was member)
> Nov 26 10:44:25 rnas2 cib[17250]: notice: Removing altai/2 from
> the membership list
> Nov 26 10:44:25 rnas2 cib[17250]: notice: Purged 1 peers with
> id=2 and/or uname=altai from the membership cache
> Nov 26 10:44:25 rnas2 pacemakerd[17249]: notice: Node altai[2]
> - state is now lost (was member)
> Nov 26 10:44:25 rnas2 crmd[17255]: notice: Node altai[2] -
> state is now lost (was member)
> Nov 26 10:44:25 rnas2 stonith-ng[17251]: notice:
> crm_update_peer_proc: Node altai[2] - state is now lost (was member)
> Nov 26 10:44:25 rnas2 crmd[17255]: warning: No match for
> shutdown action on 2
> Nov 26 10:44:25 rnas2 attrd[17253]: notice: Removing altai/2
> from the membership list
> Nov 26 10:44:25 rnas2 crmd[17255]: notice: Stonith/shutdown of
> altai not matched
> Nov 26 10:44:25 rnas2 stonith-ng[17251]: notice: Removing
> altai/2 from the membership list
> Nov 26 10:44:25 rnas2 attrd[17253]: notice: Purged 1 peers with
> id=2 and/or uname=altai from the membership cache
> Nov 26 10:44:25 rnas2 crmd[17255]: notice: State transition
> S_IDLE -> S_POLICY_ENGINE [ input=I_PE_CALC cause=C_FSA_INTERNAL
> origin=abort_transition_graph ]
> Nov 26 10:44:25 rnas2 stonith-ng[17251]: notice: Purged 1 peers
> with id=2 and/or uname=altai from the membership cache
> Nov 26 10:44:25 rnas2 crmd[17255]: warning: No match for
> shutdown action on 2
> Nov 26 10:44:25 rnas2 crmd[17255]: notice: Stonith/shutdown of
> altai not matched
> Nov 26 10:44:25 rnas2 pengine[17254]: notice: Restart nfs
> -grace:0 (Started kaukasus)
> Nov 26 10:44:25 rnas2 pengine[17254]: notice: Restart nfs
> -grace:1 (Started rnas2)
> Nov 26 10:44:25 rnas2 pengine[17254]: notice: Restart kaukasus
> -cluster_ip-1 (Started kaukasus)
> Nov 26 10:44:25 rnas2 pengine[17254]: notice: Start altai
> -cluster_ip-1 (kaukasus)
> Nov 26 10:44:25 rnas2 pengine[17254]: notice: Start altai
> -trigger_ip-1 (kaukasus)
> Nov 26 10:44:25 rnas2 pengine[17254]: notice: Restart rnas2
> -cluster_ip-1 (Started rnas2)
> Nov 26 10:44:25 rnas2 pengine[17254]: notice: Calculated
> Transition 85: /var/lib/pacemaker/pengine/pe-input-86.bz2
> Nov 26 10:44:25 rnas2 crmd[17255]: notice: Initiating action
> 29: stop kaukasus-cluster_ip-1_stop_0 on kaukasus
> Nov 26 10:44:25 rnas2 crmd[17255]: notice: Initiating action
> 35: start altai-trigger_ip-1_start_0 on kaukasus
> Nov 26 10:44:25 rnas2 crmd[17255]: notice: Initiating action
> 37: stop rnas2-cluster_ip-1_stop_0 on rnas2 (local)
> Nov 26 10:44:25 rnas2 crmd[17255]: notice: Initiating action
> 36: monitor altai-trigger_ip-1_monitor_10000 on kaukasus
> Nov 26 10:44:25 rnas2 IPaddr(rnas2-cluster_ip-1)[30797]: INFO:
> IP status = ok, IP_CIP> Nov 26 10:44:25 rnas2 crmd[17255]: notice:
Operation rnas2
> -cluster_ip-1_stop_0: ok (node=rnas2, call=82, rc=0, cib-update=210,
> confirmed=true)
> Nov 26 10:44:25 rnas2 crmd[17255]: notice: Initiating action
> 21: stop nfs-grace_stop_0 on kaukasus
> Nov 26 10:44:25 rnas2 crmd[17255]: notice: Initiating action
> 23: stop nfs-grace_stop_0 on rnas2 (local)
> Nov 26 10:44:25 rnas2 crmd[17255]: notice: Operation nfs
> -grace_stop_0: ok (node=rnas2, call=84, rc=0, cib-update=211,
> confirmed=true)
> Nov 26 10:44:25 rnas2 crmd[17255]: notice: Initiating action
> 22: start nfs-grace_start_0 on kaukasus
> Nov 26 10:44:25 rnas2 crmd[17255]: notice: Initiating action
> 24: start nfs-grace_start_0 on rnas2 (local)
> Nov 26 10:44:26 rnas2 ntpd[1700]: Deleting interface #27 bond0,
> 129.132.145.23#123, interface stats: received=0, sent=0, dropped=0,
> active_time=69258 secs
> Nov 26 10:45:05 rnas2 lrmd[17252]: warning: nfs-grace_start_0
> process (PID 30810) timed out
> Nov 26 10:45:05 rnas2 lrmd[17252]: warning: nfs
> -grace_start_0:30810 - timed out after 40000ms
> Nov 26 10:45:05 rnas2 crmd[17255]: error: Operation nfs
> -grace_start_0: Timed Out (node=rnas2, call=85, timeout=40000ms)
> Nov 26 10:45:05 rnas2 crmd[17255]: warning: Action 24 (nfs
> -grace_start_0) on rnas2 failed (target: 0 vs. rc: 1): Error
> Nov 26 10:45:05 rnas2 crmd[17255]: notice: Transition aborted
> by nfs-grace_start_0 'modify' on rnas2: Event failed
> (magic=2:1;24:85:0:836713e1-c9d3-43f8-bffd-756e023eee8a,...event:381,
> 0)
> Nov 26 10:45:05 rnas2 crmd[17255]: warning: Action 24 (nfs
> -grace_start_0) on rnas2 failed (target: 0 vs. rc: 1): Error
> Nov 26 10:45:05 rnas2 crmd[17255]: warning: Action 22 (nfs
> -grace_start_0) on kaukasus failed (target: 0 vs. rc: 1): Error
> Nov 26 10:45:05 rnas2 crmd[17255]: warning: Action 22 (nfs
> -grace_start_0) on kaukasus failed (target: 0 vs. rc: 1): Error
> Nov 26 10:45:05 rnas2 crmd[17255]: notice: Transition 85
> (Complete=13, Pending=0, Fired=0, Skipped=3, Incomplete=8,
> Source=/var/lib/pacemaker/pengine/pe-input-86.bz2): Stopped
> Nov 26 10:45:05 rnas2 pengine[17254]: warning: Processing
> failed op start for nfs-grace:0 on kaukasus: unknown error (1)
> Nov 26 10:45:05 rnas2 pengine[17254]: warning: Processing
> failed op start for nfs-grace:0 on kaukasus: unknown error (1)
> Nov 26 10:45:05 rnas2 pengine[17254]: warning: Processing
> failed op start for nfs-grace:1 on rnas2: unknown error (1)
> Nov 26 10:45:05 rnas2 pengine[17254]: warning: Processing
> failed op start for nfs-grace:1 on rnas2: unknown error (1)
> Nov 26 10:45:05 rnas2 pengine[17254]: warning: Forcing nfs
> -grace-clone away from rnas2 after 1000000 failures (max=1000000)
> Nov 26 10:45:05 rnas2 pengine[17254]: warning: Forcing nfs
> -grace-clone away from rnas2 after 1000000 failures (max=1000000)
> Nov 26 10:45:05 rnas2 pengine[17254]: warning: Forcing nfs
> -grace-clone away from rnas2 after 1000000 failures (max=1000000)
> Nov 26 10:45:05 rnas2 pengine[17254]: notice: Recover nfs
> -grace:0 (Started kaukasus)
> Nov 26 10:45:05 rnas2 pengine[17254]: notice: Stop nfs
> -grace:1 (rnas2)
> Nov 26 10:45:05 rnas2 pengine[17254]: notice: Start kaukasus
> -cluster_ip-1 (kaukasus)
> Nov 26 10:45:05 rnas2 pengine[17254]: notice: Start altai
> -cluster_ip-1 (kaukasus)
> Nov 26 10:45:05 rnas2 pengine[17254]: notice: Start rnas2
> -cluster_ip-1 (rnas2)
> Nov 26 10:45:05 rnas2 pengine[17254]: notice: Calculated
> Transition 86: /var/lib/pacemaker/pengine/pe-input-87.bz2
> Nov 26 10:45:05 rnas2 pengine[17254]: warning: Processing
> failed op start for nfs-grace:0 on kaukasus: unknown error (1)
> Nov 26 10:45:05 rnas2 pengine[17254]: warning: Processing
> failed op start for nfs-grace:0 on kaukasus: unknown error (1)
> Nov 26 10:45:05 rnas2 pengine[17254]: warning: Processing
> failed op start for nfs-grace:1 on rnas2: unknown error (1)
> Nov 26 10:45:05 rnas2 pengine[17254]: warning: Processing
> failed op start for nfs-grace:1 on rnas2: unknown error (1)
> Nov 26 10:45:05 rnas2 pengine[17254]: warning: Forcing nfs
> -grace-clone away from kaukasus after 1000000 failures (max=1000000)
> Nov 26 10:45:05 rnas2 pengine[17254]: warning: Forcing nfs
> -grace-clone away from kaukasus after 1000000 failures (max=1000000)
> Nov 26 10:45:05 rnas2 pengine[17254]: warning: Forcing nfs
> -grace-clone away from kaukasus after 1000000 failures (max=1000000)
> Nov 26 10:45:05 rnas2 pengine[17254]: warning: Forcing nfs
> -grace-clone away from rnas2 after 1000000 failures (max=1000000)
> Nov 26 10:45:05 rnas2 pengine[17254]: warning: Forcing nfs
> -grace-clone away from rnas2 after 1000000 failures (max=1000000)
> Nov 26 10:45:05 rnas2 pengine[17254]: warning: Forcing nfs
> -grace-clone away from rnas2 after 1000000 failures (max=1000000)
> Nov 26 10:45:05 rnas2 pengine[17254]: notice: Stop nfs
> -grace:0 (kaukasus)
> Nov 26 10:45:05 rnas2 pengine[17254]: notice: Stop nfs
> -grace:1 (rnas2)
> Nov 26 10:45:05 rnas2 pengine[17254]: notice: Start kaukasus
> -cluster_ip-1 (kaukasus - blocked)
> Nov 26 10:45:05 rnas2 pengine[17254]: notice: Start altai
> -cluster_ip-1 (kaukasus - blocked)
> Nov 26 10:45:05 rnas2 pengine[17254]: notice: Start rnas2
> -cluster_ip-1 (rnas2 - blocked)
> Nov 26 10:45:05 rnas2 pengine[17254]: notice: Calculated
> Transition 87: /var/lib/pacemaker/pengine/pe-input-88.bz2
> Nov 26 10:45:05 rnas2 crmd[17255]: notice: Initiating action 2:
> stop nfs-grace_stop_0 on kaukasus
> Nov 26 10:45:05 rnas2 crmd[17255]: notice: Initiating action 6:
> stop nfs-grace_stop_0 on rnas2 (local)
> Nov 26 10:45:05 rnas2 crmd[17255]: notice: Operation nfs
> -grace_stop_0: ok (node=rnas2, call=86, rc=0, cib-update=218,
> confirmed=true)
> Nov 26 10:45:05 rnas2 crmd[17255]: notice: Transition 87
> (Complete=5, Pending=0, Fired=0, Skipped=0, Incomplete=0,
> Source=/var/lib/pacemaker/pengine/pe-input-88.bz2): Complete
> Nov 26 10:45:05 rnas2 crmd[17255]: notice: State transition
> S_TRANSITION_ENGINE -> S_IDLE [ input=I_TE_SUCCESS cause=C_FSA_INTERNAL
> origin=notify_crmd ]
>
> Yours,
> Rigi
> _______________________________________________
> Gluster-users mailing list
> Gluster-users at gluster.org
> http://www.gluster.org/mailman/listinfo/gluster-users
>