[Pacemaker] Resource is not started immediately regardless of require-all="false"
Andrew Beekhof
andrew at beekhof.net
Mon Sep 23 09:27:54 CEST 2013
On 12/09/2013, at 8:29 PM, Takatoshi MATSUO <matsuo.tak at gmail.com> wrote:
> Hi all
>
> I use resource_set and require-all='false' as follows
>
> --------------------------------
> <constraints>
> <rsc_order id="order1" kind="Mandatory">
> <resource_set id="order1-0">
> <resource_ref id="dummy1"/>
> </resource_set>
> <resource_set id="order1-1" require-all="false" sequential="false">
> <resource_ref id="dummy2"/>
> <resource_ref id="dummy3"/>
> </resource_set>
> <resource_set id="order1-2">
> <resource_ref id="dummy4"/>
> </resource_set>
> </rsc_order>
> </constraints>
> --------------------------------
>
> Dummy1 and dummy2 have "sleep 5" on start.
> and dummy3 has "sleep 10" on start.
>
> In this case, dummy4 should be started immediately after dummy2 is started ?
> But dummy4 is not started until dummy3 is started.
>
> dummy1 17:50:10 + sleep 5
> dummy2 17:50:15 + sleep 5
> dummy3 17:50:15 + sleep 10
> dummy4 17:50:25 <--- not 17:50:20 !!
>
> Is this normal ?
Unfortunately yes, although I can see why you'd think it would work differently.
sequential="false" allows dummy2 and dummy3 to start in parallel
require-all="false" allows dummy4 to start if only one of dummy2 and dummy3
but if dummy2 and dummy3 are both starting, then dummy4 still has to wait for both of them.
What you're thinking off would be useful, but isn't possible yet.
>
> I use pacemaker 951cc5f which is commited 7 days ago.
>
> --------------------------------------------
> Last updated: Thu Sep 12 17:52:26 2013
> Last change: Thu Sep 12 17:50:09 2013 via cibadmin on 16-sl6
> Stack: corosync
> Current DC: 16-sl6 (1084752244) - partition with quorum
> Version: 1.1.11-0.1.951cc5f.git.el6-951cc5f
> 1 Nodes configured
> 4 Resources configured
>
>
> Online: [ 16-sl6 ]
>
> dummy1 (ocf::heartbeat:Dummy5): Started 16-sl6
> dummy2 (ocf::heartbeat:Dummy5): Started 16-sl6
> dummy3 (ocf::heartbeat:Dummy10): Started 16-sl6
> dummy4 (ocf::heartbeat:Dummy): Started 16-sl6
>
> Operations:
> * Node 16-sl6:
> dummy3: migration-threshold=1
> + (30) start: last-rc-change='Thu Sep 12 17:50:15 2013'
> last-run='Thu Sep 12 17:50:15 2013' exec-time=10050ms queue-time=0ms
> rc=0 (ok)
> + (38) monitor: interval=10000ms last-rc-change='Thu Sep 12
> 17:50:25 2013' exec-time=27ms queue-time=0ms rc=0 (ok)
> dummy4: migration-threshold=1
> + (40) start: last-rc-change='Thu Sep 12 17:50:25 2013'
> last-run='Thu Sep 12 17:50:25 2013' exec-time=40ms queue-time=0ms rc=0
> (ok)
> + (44) monitor: interval=10000ms last-rc-change='Thu Sep 12
> 17:50:25 2013' exec-time=25ms queue-time=0ms rc=0 (ok)
> dummy1: migration-threshold=1
> + (23) start: last-rc-change='Thu Sep 12 17:50:10 2013'
> last-run='Thu Sep 12 17:50:10 2013' exec-time=5036ms queue-time=0ms
> rc=0 (ok)
> + (26) monitor: interval=10000ms last-rc-change='Thu Sep 12
> 17:50:15 2013' exec-time=44ms queue-time=0ms rc=0 (ok)
> dummy2: migration-threshold=1
> + (28) start: last-rc-change='Thu Sep 12 17:50:15 2013'
> last-run='Thu Sep 12 17:50:15 2013' exec-time=5062ms queue-time=0ms
> rc=0 (ok)
> + (34) monitor: interval=10000ms last-rc-change='Thu Sep 12
> 17:50:20 2013' exec-time=23ms queue-time=0ms rc=0 (ok)
> --------------------------------------------
>
> Regards,
> Takatoshi MATSUO
>
> _______________________________________________
> Pacemaker mailing list: Pacemaker at 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
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 841 bytes
Desc: Message signed with OpenPGP using GPGMail
URL: <http://oss.clusterlabs.org/pipermail/pacemaker/attachments/20130923/da8358c6/attachment-0001.sig>
More information about the Pacemaker
mailing list