hi,

during my tests with clustered ips+drbd+ocfs2+services, i encountered a
strange problem.

i have a drbd clone:
  <master_slave id="ms_drbd_www">
    <primitive class="ocf" id="drbd_www" provider="heartbeat" type="drbd">
      ...
    </primitive>
  </master_slave>

with the rsc_order/rsc_colocation rules:

 <rsc_order id="drbd_before_ocfs2_www" action="start" from="clone_ocfs2_www" type="after" 
to_action="promote" to="ms_drbd_www" />
 <rsc_colocation id="colo_ddrbd_ocfs2_www" to="ms_drbd_www"   to_role="master" 
from="clone_ocfs2_www" score="INFINITY"/>

using "ptest -L -VVVVVV 2>&1|grep -i order|cut -d " " -f 3-|grep ocfs"
i find that the ordering is:

debug: unpack_rsc_order: order-11 (drbd_before_ocfs2_www): ms_drbd_www_promote 
before clone_ocfs2_www_start flags=0x000112
debug: unpack_rsc_order: order-12 (drbd_before_ocfs2_www): clone_ocfs2_www_stop 
before ms_drbd_www_demote flags=0x000101
debug: unpack_rsc_order: order-15 (webservices_after_ocfs2_www): 
clone_ocfs2_www_start before clone_webservice_start flags=0x000112
debug: unpack_rsc_order: order-16 (webservices_after_ocfs2_www): 
clone_webservice_stop before clone_ocfs2_www_stop flags=0x000101
ERROR: clone_rsc_order_lh: Clone-to-* ordering: clone_ocfs2_www_stop_0 -> 
ms_drbd_www_demote_0 0x000101
ERROR: clone_rsc_order_lh: Clone-to-clone ordering: clone_webservice_stop_0 -> 
clone_ocfs2_www_stop_0 0x000101

which seems perfektly right.

now my logfiles show all the first-time monitoring actions, and then
crmd[21313]: 2007/10/09_16:58:54 info: process_lrm_event: LRM operation 
extip_2_start_0 (call=18, rc=0) complete
crmd[21313]: 2007/10/09_16:58:54 info: process_lrm_event: LRM operation 
extip_1_start_0 (call=17, rc=0) complete
crmd[21313]: 2007/10/09_16:58:54 info: process_lrm_event: LRM operation 
intip_2_slave_start_0 (call=20, rc=0) complete
crmd[21313]: 2007/10/09_16:58:54 info: process_lrm_event: LRM operation 
intip_1_master_start_0 (call=19, rc=0) complete
crmd[21313]: 2007/10/09_16:58:55 ERROR: process_lrm_event: LRM operation 
ocfs2_www:1_start_0 (call=21, rc=1) Error unknown err
or
 > *monitoring actions*
crmd[21313]: 2007/10/09_16:58:57 info: process_lrm_event: LRM operation 
ocfs2_www:1_stop_0 (call=28, rc=0) complete
crmd[21313]: 2007/10/09_16:58:57 ERROR: process_lrm_event: LRM operation 
ocfs2_www:0_start_0 (call=31, rc=1) Error unknown err
or
crmd[21313]: 2007/10/09_16:58:57 info: process_lrm_event: LRM operation 
drbd_www:0_start_0 (call=29, rc=0) complete
crmd[21313]: 2007/10/09_16:58:57 info: process_lrm_event: LRM operation 
drbd_www:0_notify_0 (call=32, rc=0) complete
crmd[21313]: 2007/10/09_16:58:58 info: process_lrm_event: LRM operation 
drbd_www:0_notify_0 (call=35, rc=0) complete
 > ...
crmd[21313]: 2007/10/09_16:58:58 info: process_lrm_event: LRM operation 
ocfs2_www:0_stop_0 (call=37, rc=0) complete
 > ...
crmd[21313]: 2007/10/09_16:58:58 info: process_lrm_event: LRM operation 
drbd_www:0_promote_0 (call=38, rc=0) complete
 > ...
crmd[21313]: 2007/10/09_16:58:58 info: process_lrm_event: LRM operation 
drbd_www:0_notify_0 (call=41, rc=0) complete
 > ...
crmd[21313]: 2007/10/09_16:59:05 info: process_lrm_event: LRM operation 
ocfs2_www:0_start_0 (call=43, rc=0) complete
crmd[21313]: 2007/10/09_16:59:05 info: process_lrm_event: LRM operation 
ocfs2_www:0_monitor_120000 (call=48, rc=0) complete

any ideas, why ocfs2_www is started before drbd_www?

might it be related to a "strange" first monitoring action of drbd:
drbd[21448]:    2007/10/09_16:58:52 INFO: www: Using hostname node_0
drbd[21448]:    2007/10/09_16:58:52 DEBUG: www: Calling drbdadm -c 
/etc/drbd.conf state www
 > ...
drbd[21448]:    2007/10/09_16:58:52 DEBUG: www: Exit code 0
drbd[21448]:    2007/10/09_16:58:52 DEBUG: www: Command output:  Unconfigured
drbd[21448]:    2007/10/09_16:58:52 DEBUG: www: Calling drbdadm -c 
/etc/drbd.conf cstate www
drbd[21448]:    2007/10/09_16:58:52 DEBUG: www: Exit code 0
drbd[21448]:    2007/10/09_16:58:52 DEBUG: www: Command output:  
/etc/drbd.conf:368: in resource www, on node_0: IP 192.168.10 0.201 not found 
on this host. /etc/drbd.conf:428: in resource mysql1, on node_0: IP 
192.168.100.201 not found on this host.
drbd[21448]:    2007/10/09_16:58:52 DEBUG: www status: Unconfigured Not 
configured Not configured Unconfigured
drbd[21448]:    2007/10/09_16:58:52 DEBUG: www monitor: resource not configured
crmd[21313]: 2007/10/09_16:58:52 info: process_lrm_event: LRM operation 
drbd_www:0_monitor_0 (call=7, rc=7) complete
tengine[21328]: 2007/10/09_16:58:52 info: match_graph_event: Action 
drbd_www:0_monitor_0 (8) confirmed on webcluster01 (rc=0)

as far as i can see the drbd monitoring has the same returncode (rc) as
all the other monitoring operations (e.g. for the not-yet-started ip).

thank you for your help,
raoul bhatia
--
____________________________________________________________________
DI (FH) Raoul Bhatia M.Sc.          email.          [EMAIL PROTECTED]
Technischer Leiter

IPAX - Aloy Bhatia Hava OEG         web.          http://www.ipax.at
Barawitzkagasse 10/2/2/11           email.            [EMAIL PROTECTED]
1190 Wien                           tel.               +43 1 3670030
FN 277995t HG Wien                  fax.            +43 1 3670030 15
____________________________________________________________________

Attachment: pe-warn-243.bz2
Description: Binary data

Attachment: pe-warn-244.bz2
Description: Binary data

Attachment: pe-warn-245.bz2
Description: Binary data

Attachment: cib_running.xml.gz
Description: application/gzip

Attachment: ha-debug.gz
Description: application/gzip

_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems

Reply via email to