The pre-reqs for a functional OCFS2 cluster (using the userspace stack) are:
* pacemaker 1.0
* kernel >=2.6.26
* ocfs2-tools >= 1.4.1
* openais

Heartbeat was only a possibility for SLES10 kernels which had this
capability patched in.
The patches were not accepted upstream and have since been deprecated.
The chances of getting them to apply on a recent kernel are minimal.

On Tue, Sep 8, 2009 at 12:38 AM, Christoph Lechner<[email protected]> wrote:
> Christoph Lechner wrote:
>> Hi all,
>>
>> DRBD failover isn't working for me :(
>>
>> I'm running Heartbeat 2.1.4 with CRM enabled in an active/active setup.
>> dopd is enabled as is the drbd-peer-outdater in the drbd configuration.
>>
>> All the resources depending on DRBD are located on the host running as
>> DRBD master. If I kill the virtual machine running the DRBD master host,
>> I'm expecting the other machine to take over after some time. But
>> nothing happens, only some log messages looping over and over again, 2
>> seconds between the block of messages down below popping up again in the
>> syslog:
>>
>>> Sep  7 19:51:11 rt1 /usr/lib/heartbeat/dopd: [1998]: WARN: Cluster node: 
>>> rt2: status: dead
>>> Sep  7 19:51:11 rt1 /usr/lib/heartbeat/dopd: [1998]: debug: outdater: no 
>>> message this time
>>> Sep  7 19:51:11 rt1 /usr/lib/heartbeat/dopd: [1998]: debug: Processed 1 
>>> messages
>>> Sep  7 19:51:11 rt1 /usr/lib/heartbeat/dopd: [1998]: debug: destroying 
>>> connection: (null)
>>> Sep  7 19:51:11 rt1 /usr/lib/heartbeat/dopd: [1998]: debug: Deleting 
>>> outdater (0x8cabc88) from mainloop
>>> Sep  7 19:51:11 rt1 /usr/lib/heartbeat/dopd: [1998]: debug: Connecting 
>>> channel
>>> Sep  7 19:51:11 rt1 /usr/lib/heartbeat/dopd: [1998]: debug: Client outdater 
>>> (0x8cabc88) connected
>>> Sep  7 19:51:11 rt1 /usr/lib/heartbeat/dopd: [1998]: debug: invoked: 
>>> outdater
>>> Sep  7 19:51:11 rt1 /usr/lib/heartbeat/dopd: [1998]: debug: Processing msg 
>>> from outdater
>>> Sep  7 19:51:11 rt1 /usr/lib/heartbeat/dopd: [1998]: debug: Got message 
>>> from (drbd-peer-outdater). (peer: rt2, res :r0)
>>> Sep  7 19:51:11 rt1 /usr/lib/heartbeat/dopd: [1998]: debug: Starting node 
>>> walk
>>> Sep  7 19:51:13 rt1 kernel: [ 3706.060533] drbd0: helper command: 
>>> /sbin/drbdadm outdate-peer minor-0 exit code 20 (0x1400)
>>> Sep  7 19:51:13 rt1 kernel: [ 3706.060538] drbd0: outdate-peer helper 
>>> broken, returned 20
>>> Sep  7 19:51:13 rt1 kernel: [ 3706.060892] drbd0: helper command: 
>>> /sbin/drbdadm outdate-peer minor-0
>>> Sep  7 19:51:13 rt1 drbd-peer-outdater: [7205]: debug: message: 
>>> outdater_rc, rt1
>>> Sep  7 19:51:13 rt1 drbd-peer-outdater: [7209]: debug: drbd peer: rt2
>>> Sep  7 19:51:13 rt1 drbd-peer-outdater: [7209]: debug: drbd resource: r0
>>
>> 30 minutes now since I killed the master host.
>>
>> Where's my fault?
> The first fault that I've found was that I wasn't running Heartbeat
> 2.1.4 but still running 2.1.3 . I forgot to install the newly-built deb
> packages. I'm somewhat ashamed of mixing up the installed version number.
>
> The post
>> http://www.nabble.com/Re%3A--PATCH--dopd-should-notify-when-peer-is-dead-%28was-%22Refusing-to-be-Primary-while-peer-is-not-outdated%22-when-peer-is-dead-%29-p15738134.html
> indicated that there was something wrong with the dopd in Heartbeat
> 2.1.3 an I lookup up the version number in hb_gui ...
>
> Now failover works, but ocfs2 doesn't work anymore. Have to debug it
> tomorrow. It appears that the Filesystem OCF script in 2.1.3 mounts the
> file system, but the one shipped along with 2.1.4 doesn't. The reason
> seams to be that in Filesystem_notify the creating of a symlink in the
> o2cb config FS fails. It's in line 563, for the curious.
> The failing line of code is:
>
> if ! ln -s $OCFS2_CLUSTER_ROOT/node/$entry $OCFS2_FS_ROOT/$entry ; then
>
> My OCFS2 Kernel code has version 1.5.0, it's part of Debian stable (=5.0).
>
>  - cl
> _______________________________________________
> 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

Reply via email to