On Feb 25, 2008, at 2:16 PM, Dejan Muhamedagic wrote:
Hi,
On Fri, Feb 22, 2008 at 04:10:57PM -0600, Michael Brennen wrote:
On Friday 22 February 2008, Michael Brennen wrote:
<rsc_location id="location_iscsi106" rsc="iscsi_fs">
<rule id="prefered_location_iscsi106" score="-INFINITY"
boolean_op="and">
<expression attribute="#uname" id="location_iscsi_01"
operation="ne"
value="lvc1.fishnet.us"/>
<expression attribute="#uname" id="location_iscsi_02"
operation="ne"
value="lvc2.fishnet.us"/>
</rule>
</rsc_location>
ERROR: <expression operation="ne"...>: [ne] is not a legal value for
attribute operation
Is that correct? The above is derived from this page:
http://www.linux-ha.org/DRBD/HowTov2
<rsc_location id="drbd0-placement-1" rsc="ms-drbd0">
<rule id="drbd0-rule-1" score="-INFINITY">
<expression id="exp-01" value="xen-1" attribute="#uname"
operation="ne"/>
<expression id="exp-02" value="xen-2" attribute="#uname"
operation="ne"/>
</rule>
</rsc_location>
There is a mismatch between ciblint and heartbeat. ciblint only
acknowledges 'neq'; heartbeat only accepts 'ne'. I ran a test of
neq and
this error was thrown in the logs.
Feb 22 15:27:38 lvc4 cib: [25389]: ERROR: Value "neq" for attribute
operation
of expression is not among the enumerated set
It seems like ciblint is wrong. Please open a bug for it.
Part of the perplexing problem of the cib that I sent earlier is
that the
location directives are not followed. In particular the iscsi
resource
continually tries to run on lvc4, but per the above location
expression "location_iscsi106", it is forbidden to run on anything
except
lvc1 or lvc2. I have wondered if the location expressions were being
evaluated, and now this question of 'ne' and 'neq' is another
wrinkle. I
guess I still wonder if I don't have something fundamentally wrong.
I may try an asymmetric cluster with positively specified locations.
The latest cib is attached; the major lint complaint now is the 'ne'
expression.
The other lint complaint is 'transition_idle_timeout' in
crm_config. This is
the spelling documented in crm.dtd. This originally came from
http://wiki.linux-ha.org/AtulAthavale, where it is
spelled 'transition-idle-timeout'. The only place the latter
spelling is
found in the source tree are these files:
/usr/src/heartbeat-2.1.3/tools/haresources2cib.py.in: "transition-
idle-timeout=5min",
/usr/src/heartbeat-2.1.3/tools/haresources2cib.py: "transition-
idle-timeout=5min",
The latter (with dashes) spelling should be current.
its not
There was a bug opened for that:
http://developerbugs.linux-foundation.org/show_bug.cgi?id=1801
It's not clean whether the transition-idle-timeout is supported
pengine metadata indicates not
or if it's been replaced.
the new name is cluster-delay
Thanks,
Dejan
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems
