[ClusterLabs] can't move/migrate ressource

Ken Gaillot kgaillot at redhat.com
Thu Oct 12 10:57:07 EDT 2017


On Wed, 2017-10-11 at 13:58 +0200, Stefan Krueger wrote:
> Hello,
> 
> when i try to migrate a ressource from one server to an other (for
> example for maintenance), it don't work.
> a single ressource works fine, after that I create a group with 2
> ressources and try to move that.
> 
> my config is:
> crm conf show
> node 739272007: zfs-serv1
> node 739272008: zfs-serv2
> primitive HA_IP-Serv1 IPaddr2 \
>         params ip=172.16.101.70 cidr_netmask=16 \
>         op monitor interval=20 timeout=30 on-fail=restart nic=bond0 \
>         meta target-role=Started
> primitive HA_IP-Serv2 IPaddr2 \
>         params ip=172.16.101.74 cidr_netmask=16 \
>         op monitor interval=10s nic=bond0
> primitive nc_storage ZFS \
>         params pool=nc_storage importargs="-d /dev/disk/by-
> partlabel/"
> group compl_zfs-serv1 nc_storage HA_IP-Serv1
> location cli-prefer-HA_IP-Serv1 compl_zfs-serv1 role=Started inf:
> zfs-serv1
> location cli-prefer-HA_IP-Serv2 HA_IP-Serv2 role=Started inf: zfs-
> serv2
> location cli-prefer-compl_zfs-serv1 compl_zfs-serv1 role=Started inf:
> zfs-serv2
> location cli-prefer-nc_storage compl_zfs-serv1 role=Started inf: zfs-
> serv1

Oddly, the above constraint applies to compl_zfs-serv1 even though its
name references nc_storage. So, you have two +INFINITY constraints for
compl_zfs-serv1, for both nodes. That's not a problem, but it's almost
certainly not what you want.

If you notice, the four location constraints above start with "cli-".
When you "move" a resource, the tools actually accomplish the move by
creating a new location constraint. If you just keep running "move",
you keep piling on constraints, which might conflict with each other.
It's best to clear the old constraints (the specific command varies by
the tool you're using) before doing a new move.

> property cib-bootstrap-options: \
>         have-watchdog=false \
>         dc-version=1.1.16-94ff4df \
>         cluster-infrastructure=corosync \
>         cluster-name=debian \
>         no-quorum-policy=ignore \
>         default-resource-stickiness=100 \
>         stonith-enabled=false \
>         last-lrm-refresh=1507702403
> 
> 
> command:
> crm resource move compl_zfs-serv1 zfs-serv2
> 
> 
> pacemakerlog from zfs-serv2:
> Oct 11 13:55:58 [3556] zfs-serv2        cib:     info:
> cib_perform_op:  Diff: --- 0.106.0 2
> Oct 11 13:55:58 [3556] zfs-serv2        cib:     info:
> cib_perform_op:  Diff: +++ 0.107.0 cc224b15d0a796e040b026b7c2965770
> Oct 11 13:55:58 [3556] zfs-serv2        cib:     info:
> cib_perform_op:  --
> /cib/configuration/constraints/rsc_location[@id='cli-prefer-
> compl_zfs-serv1']
> Oct 11 13:55:58 [3556] zfs-serv2        cib:     info:
> cib_perform_op:  +  /cib:  @epoch=107
> Oct 11 13:55:58 [3556] zfs-serv2        cib:     info:
> cib_process_request:     Completed cib_delete operation for section
> constraints: OK (rc=0, origin=zfs-serv1/crm_resource/3,
> version=0.107.0)
> Oct 11 13:55:58 [3561] zfs-serv2       crmd:     info:
> abort_transition_graph:  Transition aborted by deletion of
> rsc_location[@id='cli-prefer-compl_zfs-serv1']: Configuration change
> | cib=0.107.0 source=te_update_diff:444
> path=/cib/configuration/constraints/rsc_location[@id='cli-prefer-
> compl_zfs-serv1'] complete=true
> Oct 11 13:55:58 [3561] zfs-serv2       crmd:   notice:
> do_state_transition:     State transition S_IDLE -> S_POLICY_ENGINE |
> input=I_PE_CALC cause=C_FSA_INTERNAL origin=abort_transition_graph
> Oct 11 13:55:58 [3556] zfs-serv2        cib:     info:
> cib_perform_op:  Diff: --- 0.107.0 2
> Oct 11 13:55:58 [3556] zfs-serv2        cib:     info:
> cib_perform_op:  Diff: +++ 0.108.0 (null)
> Oct 11 13:55:58 [3556] zfs-serv2        cib:     info:
> cib_perform_op:  +  /cib:  @epoch=108
> Oct 11 13:55:58 [3556] zfs-serv2        cib:     info:
> cib_perform_op:  ++ /cib/configuration/constraints:  <rsc_location
> id="cli-prefer-compl_zfs-serv1" rsc="compl_zfs-serv1" role="Started"
> node="zfs-serv2" score="INFINITY"/>
> Oct 11 13:55:58 [3556] zfs-serv2        cib:     info:
> cib_process_request:     Completed cib_modify operation for section
> constraints: OK (rc=0, origin=zfs-serv1/crm_resource/4,
> version=0.108.0)
> Oct 11 13:55:58 [3561] zfs-serv2       crmd:     info:
> abort_transition_graph:  Transition aborted by rsc_location.cli-
> prefer-compl_zfs-serv1 'create': Configuration change | cib=0.108.0
> source=te_update_diff:444 path=/cib/configuration/constraints
> complete=true
> Oct 11 13:55:58 [3560] zfs-serv2    pengine:   notice:
> unpack_config:   On loss of CCM Quorum: Ignore
> Oct 11 13:55:58 [3560] zfs-serv2    pengine:     info:
> determine_online_status: Node zfs-serv2 is online
> Oct 11 13:55:58 [3560] zfs-serv2    pengine:     info:
> determine_online_status: Node zfs-serv1 is online
> Oct 11 13:55:58 [3560] zfs-serv2    pengine:     info:
> determine_op_status:     Operation monitor found resource nc_storage
> active on zfs-serv2
> Oct 11 13:55:58 [3560] zfs-serv2    pengine:     info:
> native_print:    HA_IP-
> Serv2     (ocf::heartbeat:IPaddr2):       Started zfs-serv2
> Oct 11 13:55:58 [3560] zfs-serv2    pengine:     info:
> group_print:      Resource Group: compl_zfs-serv1
> Oct 11 13:55:58 [3560] zfs-serv2    pengine:     info:
> native_print:         nc_storage (ocf::heartbeat:ZFS):   Started zfs-
> serv1
> Oct 11 13:55:58 [3560] zfs-serv2    pengine:     info:
> native_print:         HA_IP-
> Serv1        (ocf::heartbeat:IPaddr2):       Started zfs-serv1
> Oct 11 13:55:58 [3560] zfs-serv2    pengine:     info:
> LogActions:      Leave   HA_IP-Serv2     (Started zfs-serv2)
> Oct 11 13:55:58 [3560] zfs-serv2    pengine:     info:
> LogActions:      Leave   nc_storage      (Started zfs-serv1)
> Oct 11 13:55:58 [3560] zfs-serv2    pengine:     info:
> LogActions:      Leave   HA_IP-Serv1     (Started zfs-serv1)
> Oct 11 13:55:58 [3560] zfs-serv2    pengine:   notice:
> process_pe_message:      Calculated transition 8, saving inputs in
> /var/lib/pacemaker/pengine/pe-input-1348.bz2
> Oct 11 13:55:58 [3561] zfs-serv2       crmd:     info:
> handle_response: pe_calc calculation pe_calc-dc-1507722958-42 is
> obsolete
> Oct 11 13:55:58 [3556] zfs-serv2        cib:     info:
> cib_file_backup: Archived previous version as
> /var/lib/pacemaker/cib/cib-71.raw
> Oct 11 13:55:58 [3556] zfs-serv2        cib:     info:
> cib_file_write_with_digest:      Wrote version 0.107.0 of the CIB to
> disk (digest: 5471544275a5deb060ac698cd86b1ea3)
> Oct 11 13:55:59 [3560] zfs-serv2    pengine:   notice:
> unpack_config:   On loss of CCM Quorum: Ignore
> Oct 11 13:55:59 [3560] zfs-serv2    pengine:     info:
> determine_online_status: Node zfs-serv2 is online
> Oct 11 13:55:59 [3560] zfs-serv2    pengine:     info:
> determine_online_status: Node zfs-serv1 is online
> Oct 11 13:55:59 [3560] zfs-serv2    pengine:     info:
> determine_op_status:     Operation monitor found resource nc_storage
> active on zfs-serv2
> Oct 11 13:55:59 [3560] zfs-serv2    pengine:     info:
> native_print:    HA_IP-
> Serv2     (ocf::heartbeat:IPaddr2):       Started zfs-serv2
> Oct 11 13:55:59 [3560] zfs-serv2    pengine:     info:
> group_print:      Resource Group: compl_zfs-serv1
> Oct 11 13:55:59 [3560] zfs-serv2    pengine:     info:
> native_print:         nc_storage (ocf::heartbeat:ZFS):   Started zfs-
> serv1
> Oct 11 13:55:59 [3560] zfs-serv2    pengine:     info:
> native_print:         HA_IP-
> Serv1        (ocf::heartbeat:IPaddr2):       Started zfs-serv1
> Oct 11 13:55:59 [3560] zfs-serv2    pengine:     info:
> LogActions:      Leave   HA_IP-Serv2     (Started zfs-serv2)
> Oct 11 13:55:59 [3560] zfs-serv2    pengine:     info:
> LogActions:      Leave   nc_storage      (Started zfs-serv1)
> Oct 11 13:55:59 [3560] zfs-serv2    pengine:     info:
> LogActions:      Leave   HA_IP-Serv1     (Started zfs-serv1)
> Oct 11 13:55:59 [3560] zfs-serv2    pengine:   notice:
> process_pe_message:      Calculated transition 9, saving inputs in
> /var/lib/pacemaker/pengine/pe-input-1349.bz2
> Oct 11 13:55:59 [3561] zfs-serv2       crmd:     info:
> do_state_transition:     State transition S_POLICY_ENGINE ->
> S_TRANSITION_ENGINE | input=I_PE_SUCCESS cause=C_IPC_MESSAGE
> origin=handle_response
> Oct 11 13:55:59 [3561] zfs-serv2       crmd:     info:
> do_te_invoke:    Processing graph 9 (ref=pe_calc-dc-1507722958-43)
> derived from /var/lib/pacemaker/pengine/pe-input-1349.bz2
> Oct 11 13:55:59 [3561] zfs-serv2       crmd:   notice:
> run_graph:       Transition 9 (Complete=0, Pending=0, Fired=0,
> Skipped=0, Incomplete=0, Source=/var/lib/pacemaker/pengine/pe-input-
> 1349.bz2): Complete
> Oct 11 13:55:59 [3561] zfs-serv2       crmd:     info: do_log:  Input
> I_TE_SUCCESS received in state S_TRANSITION_ENGINE from notify_crmd
> Oct 11 13:55:59 [3561] zfs-serv2       crmd:   notice:
> do_state_transition:     State transition S_TRANSITION_ENGINE ->
> S_IDLE | input=I_TE_SUCCESS cause=C_FSA_INTERNAL origin=notify_crmd
> Oct 11 13:55:59 [3556] zfs-serv2        cib:     info:
> cib_file_write_with_digest:      Reading cluster configuration file
> /var/lib/pacemaker/cib/cib.p8bDAb (digest:
> /var/lib/pacemaker/cib/cib.wWbjqa)
> Oct 11 13:55:59 [3556] zfs-serv2        cib:     info:
> cib_file_backup: Archived previous version as
> /var/lib/pacemaker/cib/cib-72.raw
> Oct 11 13:55:59 [3556] zfs-serv2        cib:     info:
> cib_file_write_with_digest:      Wrote version 0.108.0 of the CIB to
> disk (digest: 01ea463828a6de1cfcf7b14e67a0dd67)
> Oct 11 13:55:59 [3556] zfs-serv2        cib:     info:
> cib_file_write_with_digest:      Reading cluster configuration file
> /var/lib/pacemaker/cib/cib.kA8iQp (digest:
> /var/lib/pacemaker/cib/cib.Va05np)
> Oct 11 13:56:03 [3556] zfs-serv2        cib:     info:
> cib_process_ping:        Reporting our current digest to zfs-serv2:
> 10209a8d0494a10c01a2bab61691541a for 0.108.0 (0x55dcc114c920 0)
> 
> 
> thanks in advance
> 
> _______________________________________________
> Users mailing list: Users at clusterlabs.org
> http://lists.clusterlabs.org/mailman/listinfo/users
> 
> Project Home: http://www.clusterlabs.org
> Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.
> pdf
> Bugs: http://bugs.clusterlabs.org
-- 
Ken Gaillot <kgaillot at redhat.com>




More information about the Users mailing list