Hi guys,
I've just had something odd happen to me and I'm not totally sure why.
This is the result of showscores.sh on cleanly bringing up both nodes.
drbd0 and drbd1 are masterslave resources, pingd is a clone resource
and the rest are split into two groups with a Filesystem_2 and
Filesystem_12 being the first primitive in each.
drbd0 and drbd1 have resource stickiness set to 1000, as does pingd.
Basically I wanted to allow the servers to be failed over many, many
time before needing to reset scores.
Score Resource Node Stickiness
Failcount Failure-Stickiness
drbd0:0 0 kodiak.domain.org100 0
-10
drbd0:0 1075 polar.domain.org100 0
-10
drbd0:1 1075 kodiak.domain.org100 0
-10
drbd0:1 -INFINITY polar.domain.org100 0
-10
drbd1:0 0 kodiak.domain.org100 0
-10
drbd1:0 1075 polar.domain.org100 0
-10
drbd1:1 1075 kodiak.domain.org100 0
-10
drbd1:1 -INFINITY polar.domain.org100 0
-10
Filesystem_12 1475 polar.domain.org100 0
-10
Filesystem_2 1475 kodiak.domain.org100 0
-10
IPaddr_10_0_7_183INFINITY kodiak.domain.org100 0
-10
IPaddr_10_0_7_183-INFINITY polar.domain.org100 0
-10
IPaddr_10_0_7_184-INFINITY kodiak.domain.org100 0
-10
IPaddr_10_0_7_184INFINITY polar.domain.org100 0
-10
nfs-common_14 -INFINITY kodiak.domain.org100 0
-10
nfs-common_14 INFINITY polar.domain.org100 0
-10
nfs-common_4 INFINITY kodiak.domain.org100 0
-10
nfs-common_4 -INFINITY polar.domain.org100 0
-10
nfs-kernel-server_15-INFINITY kodiak.domain.org100 0
-10
nfs-kernel-server_15INFINITY polar.domain.org100 0
-10
nfs-kernel-server_5INFINITY kodiak.domain.org100 0
-10
nfs-kernel-server_5-INFINITY polar.domain.org100 0
-10
pingd-child:0 0 kodiak.domain.org100 0
-10
pingd-child:0 1000 polar.domain.org100 0
-10
pingd-child:1 1000 kodiak.domain.org100 0
-10
pingd-child:1 -INFINITY polar.domain.org100 0
-10
I was firstly wondering why the two Filesystem's don't show scores for
both nodes? I also had a problem when I rebooted one of the servers.
These servers boot up pretty fast, so I'm wondering whether a large
monitoring gap and a fast reboot can lead to problems?
The drbd resource on the rebooted node transferred as expected, as did
the Filesystem, but NOT the IP. The IP previously from the rebooted
server was set to -INFINITY for both nodes and I'm not sure why.
Below are my constraints as they are during a fresh start of the
cluster. After they're up I delete the first two location constraints
as I don't want drbd resources automatically failing back once the
cluster is up.
<constraints>
<rsc_location id="loc:r0_likes_kodiak" rsc="ms-drbd0">
<rule id="rule:r0_likes_kodiak" role="master" score="1">
<expression attribute="#uname" operation="eq"
value="kodiak.domain.org" id="3ac0cc2b-e503-4b36-a960-a07ed8447b71"/>
</rule>
</rsc_location>
<rsc_location id="loc:r1_likes_polar" rsc="ms-drbd1">
<rule id="rule:r1_likes_polar" role="master" score="1">
<expression attribute="#uname" operation="eq"
value="polar.domain.org" id="5857d79c-89df-4e8b-836c-cd43c0e7ec6d"/>
</rule>
</rsc_location>
<rsc_order id="r0_before_group_1" from="group_1" action="start"
to="ms-drbd0" to_action="promote"/>
<rsc_colocation id="group_1_on_r0" to="ms-drbd0"
to_role="master" from="group_1" score="infinity"/>
<rsc_order id="r1_before_group_11" from="group_11"
action="start" to="ms-drbd1" to_action="promote"/>
<rsc_colocation id="group_11_on_r1" to="ms-drbd1"
to_role="master" from="group_11" score="infinity"/>
<rsc_location id="ms-drbd0:connected" rsc="ms-drbd0">
<rule role="master" id="ms-drbd0:connected:rule"
score="-INFINITY" boolean_op="or">
<expression id="ms-drbd0:connected:expr:undefined"
attribute="pingd" operation="not_defined"/>
<expression id="ms-drbd0:connected:expr:zero"
attribute="pingd" operation="lte" value="0"/>
</rule>
</rsc_location>
<rsc_location id="ms-drbd1:connected" rsc="ms-drbd1">
<rule role="master" id="ms-drbd1:connected:rule"
score="-INFINITY" boolean_op="or">
<expression id="ms-drbd1:connected:expr:undefined"
attribute="pingd" operation="not_defined"/>
<expression id="ms-drbd1:connected:expr:zero"
attribute="pingd" operation="lte" value="0"/>
</rule>
</rsc_location>
</constraints>
Any ideas on where I can start looking would be appreciated.
Thanks
Guy
--
Don't just do something...sit there!
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems
