[Linux-HA] Cloned resources starting on forbidden nodes

[Max Hofer]

I have a cluster with 4 nodes: management1, management2, routing1, routing2.

I created a cloned resource called "isdn_cards" (the primitive is "isdn_card") 
which should only run on "management[12]" and never on routing[12].

 The constraint in the cib.xml looks like:

<rsc_location id="isdn_location" rsc="isdn_cards">
  <!-- make sure they run only on management servers -->
  <rule id="isdn_not_on_routing_servers" score="-INFINITY" boolean_op="or">
    <expression attribute="#id" operation="eq" value="routing1"/>
    <expression attribute="#id" operation="eq" value="routing2"/>
  </rule>
  <rule id="isdn_on_mgmt_servers" score="INFINITY" boolean_op="or">
     <expression attribute="#id" operation="eq" value="management1"/>
     <expression attribute="#id" operation="eq" value="management2"/>
  </rule>
</rsc_location>

At the first it seemed to work pretty fine until i shut down all 4 nodes 
simultaneously and i restarted them at the same time.

Then the PE moves "isdn_card:1" to routing1 and i have no clue why.

Excerpt from the ha-log:

pengine[21020]: 2006/10/28_10:27:38 notice: NoRoleChange:native.c Leave  resource pingd-child:3  (routing1)
pengine[21020]: 2006/10/28_10:27:38 notice: NoRoleChange:native.c Move  resource isdn_card:0    (management1 -> routing1)
pengine[21020]: 2006/10/28_10:27:38 notice: Recurring:native.c routing1     isdn_card:0_monitor_10000
pengine[21020]: 2006/10/28_10:27:38 notice: NoRoleChange:native.c Leave resource isdn_card:1    (management2)
pengine[21020]: 2006/10/28_10:27:38 notice: NoRoleChange:native.c Leave resource clusterIp_1    (routing1)

The RA for the ISDN cards (pt-isdn) is "buggy" in the sense that stop does not
work on a node where the software for the ISDN card is not installed and thus
the resource remains in failed state (will be fixed soon - i just never thought a
start/stop would occur on routingX because of my constraints) ---
appart from that --- Why is the pengine moving the isdn_card:0 to routing1?

Sadly I never was able to figure out how ptest works *manual needed by someone
who knows how to work with it*

Attched the cib.xml and the ha-log
-- 
Max Hofer
APUS Software G.m.b.H.
A-8074 Raaba, Bahnhofstraße 1/1
T| +43 316 401629 11
F| +43 316 401629 9
W| www.apus.co.at
E| max.hofer at apus.co.at
-------------- next part --------------
A non-text attachment was scrubbed...
Name: cib-bug-report.xml.bz2
Type: application/x-bzip2
Size: 6846 bytes
Desc: not available
Url : http://lists.community.tummy.com/pipermail/linux-ha/attachments/20061030/f8f2bd08/cib-bug-report.xml-0001.bin
-------------- next part --------------
A non-text attachment was scrubbed...
Name: ha-log-bug-report.bz2
Type: application/x-bzip2
Size: 31938 bytes
Desc: not available
Url : http://lists.community.tummy.com/pipermail/linux-ha/attachments/20061030/f8f2bd08/ha-log-bug-report-0001.bin