Hello,
I'm trying to configure a cluster based in pacemaker and corosync in
two ubuntu precise servers. The cluster is for an active/standby
pop/imap server with a shared storage accesed through fibrechannel.
In order to avoid concurrent access to this shared storage, I need clvm
(maybe I'm wrong), so I'm trying to configure it. According to different
guides and howtos I have found I have configured a DLM and clvm resource:
root@myotis51:/etc/cluster# crm configure show
node myotis51
node myotis52
primitive clvm ocf:lvm2:clvmd \
params daemon_timeout="30" \
meta target-role="Started"
primitive dlm ocf:pacemaker:controld \
meta target-role="Started"
group dlm-clvm dlm clvm
clone dlm-clvm-clone dlm-clvm \
meta interleave="true" ordered="true"
property $id="cib-bootstrap-options" \
dc-version="1.1.6-9971ebba4494012a93c03b40a2c58ec0eb60f50c" \
cluster-infrastructure="cman" \
expected-quorum-votes="2" \
stonith-enabled="false" \
no-quorum-policy="ignore" \
last-lrm-refresh="1363957949"
rsc_defaults $id="rsc-options" \
resource-stickiness="100"
With this configuration, resources are not launched, I think because
DLM is failing because it's trying to launch dlm_controld.pcmk which is
not installed in my system
Mar 22 13:52:57 myotis51 pengine: [2989]: notice: LogActions: Start
dlm:0 (myotis51)
Mar 22 13:52:57 myotis51 pengine: [2989]: notice: LogActions: Leave
dlm:1 (Stopped)
Mar 22 13:52:57 myotis51 crmd: [2990]: info: te_rsc_command: Initiating
action 4: monitor dlm:0_monitor_0 on myotis51 (local)
Mar 22 13:52:57 myotis51 crmd: [2990]: info: do_lrm_rsc_op: Performing
key=4:0:7:71fa2334-a3f3-4c01-a000-7e702a32d0e2 op=dlm:0_monitor_0 )
Mar 22 13:52:57 myotis51 lrmd: [2987]: info: rsc:dlm:0 probe[2] (pid 3050)
Mar 22 13:52:57 myotis51 controld[3050]: ERROR: Setup problem: couldn't
find command: dlm_controld.pcmk
Mar 22 13:52:57 myotis51 lrmd: [2987]: info: operation monitor[2] on
dlm:0 for client 2990: pid 3050 exited with return code 5
Mar 22 13:52:57 myotis51 crmd: [2990]: info: process_lrm_event: LRM
operation dlm:0_monitor_0 (call=2, rc=5, cib-update=27, confirmed=true)
not installed
Mar 22 13:52:57 myotis51 crmd: [2990]: WARN: status_from_rc: Action 4
(dlm:0_monitor_0) on myotis51 failed (target: 7 vs. rc: 5): Error
Mar 22 13:52:57 myotis51 crmd: [2990]: info: abort_transition_graph:
match_graph_event:277 - Triggered transition abort (complete=0,
tag=lrm_rsc_op, id=dlm:0_last_failure_0,
magic=0:5;4:0:7:71fa2334-a3f3-4c01-a000-7e702a32d0e2, cib=0.32.14) :
Event failed
Mar 22 13:52:57 myotis51 crmd: [2990]: info: match_graph_event: Action
dlm:0_monitor_0 (4) confirmed on myotis51 (rc=4)
Mar 22 13:52:57 myotis51 pengine: [2989]: notice: unpack_rsc_op: Hard
error - dlm:0_last_failure_0 failed with rc=5: Preventing dlm-clvm-clone
from re-starting on myotis51
Mar 22 13:52:57 myotis51 pengine: [2989]: notice: LogActions: Leave
dlm:0 (Stopped)
Mar 22 13:52:57 myotis51 pengine: [2989]: notice: LogActions: Leave
dlm:1 (Stopped)
The problem with this is that I can't find any dlm_controld.pcmk binary
for ubuntu. Any idea on how to fix this?
The closest command I have found is dlm_controld provided with cman
packages, but then I have to replace corosync with cman. Doing this is
not a big problem for me. The fact is that I'm newbie in HA and the use
of corosync instead of cman is because this is the one documented with
pacemaker
(http://clusterlabs.org/doc/en-US/Pacemaker/1.1-plugin/html/Clusters_from_Scratch/index.html).
Is it corosync supposed to be better (or more open or more standard
based) than cman? In case corosync is more recommended, then what is the
solution for the dlm problem?
Thanks in advanced.
--
Angel L. Mateo Martínez
Sección de Telemática
Área de Tecnologías de la Información
y las Comunicaciones Aplicadas (ATICA)
http://www.um.es/atica
Tfo: 868889150
Fax: 868888337
_______________________________________________
Pacemaker mailing list: Pacemaker@oss.clusterlabs.org
http://oss.clusterlabs.org/mailman/listinfo/pacemaker
Project Home: http://www.clusterlabs.org
Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
Bugs: http://bugs.clusterlabs.org
