Re: More than one drbd resource possible in pacemaker?

On Fri, Apr 30, 2010 at 9:46 AM, Gianluca Cecchi
<gianluca.cecchi <at> gmail.com> wrote:
> Hello,
> I have configured a drbd0 resource (nfsdata) in pacemaker, acting as
> active/passive, using the linbit resource agent with master/slave config.
> It works ok in different operations I tried with pacemaker.
>
> Then on both two nodes I'm going to test ocfs2 for another drbd1 resource I
> have created (ocfs2data).
>
> Env for drbd seems ok on both nodes, with ocfs2 fs mounted on both, but the
> drbd0 pacemaker resource failed
>
> [root <at> ha1 ~]# cat /proc/drbd
> version: 8.3.6 (api:88/proto:86-91)
> GIT-hash: f3606c47cc6fcf6b3f086e425cb34af8b7a81bbf build by root <at> ha1,
> 2010-04-28 09:01:04
>  0: cs:Connected ro:Primary/Secondary ds:UpToDate/UpToDate C r----
>    ns:3128688 nr:8 dw:1027576 dr:2586722 al:524 bm:130 lo:0 pe:0 ua:0 ap:0
> ep:1 wo:d oos:0
>  1: cs:Connected ro:Primary/Primary ds:UpToDate/UpToDate C r----
>    ns:1280472 nr:3146699 dw:4427171 dr:28918 al:343 bm:129 lo:0 pe:0 ua:0
> ap:0 ep:1 wo:d oos:0
>
> [root <at> ha1 ~]# crm_mon -1
> ============
> Last updated: Fri Apr 30 09:46:28 2010
> Stack: openais
> Current DC: ha1 - partition with quorum
> Version: 1.0.8-9881a7350d6182bae9e8e557cf20a3cc5dac3ee7

Re: o2cb pacemaker agent and

On Fri, Apr 30, 2010 at 4:43 PM, Gianluca Cecchi
<gianluca.cecchi <at> gmail.com> wrote:
> Hello,
> on rh el 5.5 trying to configure ocfs2 1.4 with pacemaker 1.0.8.
> It seems I have some problems with programs/kernel modules missing.
>
> I downloaded rpm for pacemaker from clusterlabs repo and rpm for ocfs2 from
> Oracle repo:
> ocfs2-tools-1.4.4-1.el5
> ocfs2-2.6.18-194.el5-1.4.7-1.el5
> ocfs2console-1.4.4-1.el5
>
>
> In pacemaker resource agent o2cb (part of pacemaker-1.0.8-5.el5 rpm) I have:
> ...
> : ${OCF_RESKEY_stack:="pcmk"}
> : ${OCF_RESKEY_sysfs:="/sys/fs"}
> : ${OCF_RESKEY_configfs:="/sys/kernel/config"}
> : ${OCF_RESKEY_daemon_timeout:="10"} # How long to wait for things to start
> : ${OCF_RESKEY_CRM_meta_gloablly_unique:="false"}
>
> DAEMON="/usr/sbin/ocfs2_controld.${OCF_RESKEY_stack}"
> CLUSTER_STACK_FILE="${OCF_RESKEY_sysfs}/ocfs2/cluster_stack"
> LOADED_PLUGINS_FILE="${OCF_RESKEY_sysfs}/ocfs2/loaded_cluster_plugins"
> ...
>
> and
> ...
> o2cb_start() {
>

Re: MySQL and 4 instances

On Thu, Apr 29, 2010 at 7:37 PM, mike <mgbutler <at> nbnet.nb.ca> wrote:
> Hello all,
>
> We had a simple 2 node MySQL cluster - nothing special. One instance
> that worked perfectly. We recently added 3 instances and now we're
> having some issues. The problem is that Heartbeat issues a MySQL Status
> immediately after the MySQL Start .. and of course the MySQL Status will
> fail given that the instances are still coming up .

The start operation should not return until MySQL is fully up.
Is this an LSB or OCF script?

> Then I see a MySQL
> stop, followed by another start and then a status. So we get into this
> MySQL Status / Stop / Start loop and it's not pretty.  Is there a
> setting somewhere that I can change that will allow MySQL to do a start,
> wait fro say 30 seconds and THEN issue the status?
>
> Thanks
> _______________________________________________
> Linux-HA mailing list
> Linux-HA <at> lists.linux-ha.org
> http://lists.linux-ha.org/mailman/listinfo/linux-ha
> See also: http://linux-ha.org/ReportingProblems
>
_______________________________________________
Linux-HA mailing list
Linux-HA <at> lists.linux-ha.org
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems

Re: o2cb pacemaker agent and

On Mon, May 3, 2010 at 9:22 AM, Andrew Beekhof <andrew <at> beekhof.net> wrote:

> [snip]
> You would need to rebuild ocfs2-tools with pacemaker support turned on.
>
> Hmm, thaks for answering.
I did have the same idea/impression, but at some point I was stuck with
configure script complaining about
-ldlmcontrol
library missing and I was not able to find it on rh el 5 platform...
any hint about switches to give ?
Gianluca
_______________________________________________
Linux-HA mailing list
Linux-HA <at> lists.linux-ha.org
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems

Re: o2cb pacemaker agent and

On Mon, May 3, 2010 at 10:23 AM, Gianluca Cecchi
<gianluca.cecchi <at> gmail.com> wrote:
> On Mon, May 3, 2010 at 9:22 AM, Andrew Beekhof <andrew <at> beekhof.net> wrote:
>
>> [snip]
>> You would need to rebuild ocfs2-tools with pacemaker support turned on.
>>
>> Hmm, thaks for answering.
> I did have the same idea/impression, but at some point I was stuck with
> configure script complaining about
> -ldlmcontrol

I think you need the RHCS "cluster" development packages for that

> library missing and I was not able to find it on rh el 5 platform...
> any hint about switches to give ?

I think it will do "the right thing" as long as you have the
pacemaker-libs-devel installed.
_______________________________________________
Linux-HA mailing list
Linux-HA <at> lists.linux-ha.org
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems

Re: o2cb pacemaker agent and

On Mon, May 3, 2010 at 10:51 AM, Andrew Beekhof <andrew <at> beekhof.net> wrote:

> On Mon, May 3, 2010 at 10:23 AM, Gianluca Cecchi
> <gianluca.cecchi <at> gmail.com> wrote:
> > On Mon, May 3, 2010 at 9:22 AM, Andrew Beekhof <andrew <at> beekhof.net>
> wrote:
> >
> >> [snip]
> >> You would need to rebuild ocfs2-tools with pacemaker support turned on.
> >>
> >> Hmm, thaks for answering.
> > I did have the same idea/impression, but at some point I was stuck with
> > configure script complaining about
> > -ldlmcontrol
>
> I think you need the RHCS "cluster" development packages for that
>
>
I thought this too, but I manually ran a sort of

rpm -qpl *.rpm | grep dlmcontrol

in every Server, VT, Cluster, ClusterStorage of a rhel 5.5 i386 DVD, without
finding anything about this library....

Gianluca
_______________________________________________
Linux-HA mailing list
Linux-HA <at> lists.linux-ha.org
http://lists.linux-ha.org/mailman/listinfo/linux-ha

Re: MySQL and 4 instances

On 2010-05-03 09:24, Andrew Beekhof wrote:
> On Thu, Apr 29, 2010 at 7:37 PM, mike <mgbutler <at> nbnet.nb.ca> wrote:
>> Hello all,
>>
>> We had a simple 2 node MySQL cluster - nothing special. One instance
>> that worked perfectly. We recently added 3 instances and now we're
>> having some issues. The problem is that Heartbeat issues a MySQL Status
>> immediately after the MySQL Start .. and of course the MySQL Status will
>> fail given that the instances are still coming up .
> 
> The start operation should not return until MySQL is fully up.
> Is this an LSB or OCF script?

I would almost certainly assume LSB. Or even some home-grown script. The
OCF RA loops on status during start, so start doesn't return until MySQL
is definitely up.

Mike, please use the ocf:heartbeat:mysql RA bundled with the
resource-agents package.

Cheers,
Florian

_______________________________________________
Linux-HA mailing list
Linux-HA <at> lists.linux-ha.org
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems

Re: MySQL and 4 instances

Florian Haas wrote:
> On 2010-05-03 09:24, Andrew Beekhof wrote:
>   
>> On Thu, Apr 29, 2010 at 7:37 PM, mike <mgbutler <at> nbnet.nb.ca> wrote:
>>     
>>> Hello all,
>>>
>>> We had a simple 2 node MySQL cluster - nothing special. One instance
>>> that worked perfectly. We recently added 3 instances and now we're
>>> having some issues. The problem is that Heartbeat issues a MySQL Status
>>> immediately after the MySQL Start .. and of course the MySQL Status will
>>> fail given that the instances are still coming up .
>>>       
>> The start operation should not return until MySQL is fully up.
>> Is this an LSB or OCF script?
>>     
>
> I would almost certainly assume LSB. Or even some home-grown script. The
> OCF RA loops on status during start, so start doesn't return until MySQL
> is definitely up.
>
> Mike, please use the ocf:heartbeat:mysql RA bundled with the
> resource-agents package.
>
> Cheers,
> Florian
>
>   
> ------------------------------------------------------------------------
>

Re: o2cb pacemaker agent and

On Mon, May 3, 2010 at 10:57 AM, Gianluca Cecchi
<gianluca.cecchi <at> gmail.com> wrote:
> On Mon, May 3, 2010 at 10:51 AM, Andrew Beekhof <andrew <at> beekhof.net> wrote:
>
>> On Mon, May 3, 2010 at 10:23 AM, Gianluca Cecchi
>> <gianluca.cecchi <at> gmail.com> wrote:
>> > On Mon, May 3, 2010 at 9:22 AM, Andrew Beekhof <andrew <at> beekhof.net>
>> wrote:
>> >
>> >> [snip]
>> >> You would need to rebuild ocfs2-tools with pacemaker support turned on.
>> >>
>> >> Hmm, thaks for answering.
>> > I did have the same idea/impression, but at some point I was stuck with
>> > configure script complaining about
>> > -ldlmcontrol
>>
>> I think you need the RHCS "cluster" development packages for that
>>
>>
> I thought this too, but I manually ran a sort of
>
> rpm -qpl *.rpm | grep dlmcontrol
>
> in every Server, VT, Cluster, ClusterStorage of a rhel 5.5 i386 DVD, without
> finding anything about this library....

Actually I don't think its a library anymore, just a binary.
And the pacemaker integration wont be in the 5.5 packages, you'll need
to rebuild the cman 3.x source packages.

autofailback off

who knows how to keep a bonded interface slave from failing back if the
failed interface comes back up.  My problem is an intermittent sort of
failure NIC like it fails over and then comes back...  When it comes back
the switch wants to do about 60 seconds of negotiation and delay that I
cannot get my network guys to turn off, like PortFast enable.  Too hard for
them...   So I see an autofail back setting but it looks only relevant to
Cluster failover.  I just have a  bonded interface that I want to stay
failed until manual intervention tells it otherwise.

anyone?
thanks
--

-- 
View this message in context: http://old.nabble.com/autofailback-off-tp28439670p28439670.html
Sent from the Linux-HA mailing list archive at Nabble.com.

_______________________________________________
Linux-HA mailing list
Linux-HA <at> lists.linux-ha.org
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems