Login | Register For Free | Help
Search for: (Advanced)

Mailing List Archive: Linux-HA: Pacemaker

Resource is monitored only till promote... bug or config error?

  

 
Linux-HA pacemaker RSS feed   Index | Next | Previous | View Threaded


AStallmann at CONET

Jun 28, 2012, 8:05 AM

Post #1 of 9 (426 views)
Permalink
Resource is monitored only till promote... bug or config error?
Hi!

For some yet unknown reason, pacemaker does monitor the tomcat_ms resource on my dedicated master node (ucrs1) only until it gets promoted to Master. After that, there’s no further monitoring for tomcat on the master, while the slave (ucrs2) is still monitored.

This leads to some problems: If tomcat on ucrs1 dies, the cluster will never detect this and a failover won’t happen.

Here’s my configuration for tomcat:

primitive tomcat_res ocf:ucrs:tomcat \
params java_home="/usr/lib/jvm/java-7-openjdk-amd64/jre/" catalina_home="/usr/share/tomcat7" catalina_base="/var/lib/tomcat7" tomcat_user="tomcat7" java_opts=”...” \
op start interval="0" timeout="60s" \
op stop interval="0" timeout="120s" \
op monitor interval="10s" role="Master" timeout="30s" \
op monitor interval="20s" role="Slave" timeout="30s"
ms ms_tomcat tomcat_res \
meta master-max="1" master-node-max="1" clone-max="2" clone-node-max="1" target-role="Started"

Strangely, when I ask crm what it thinks about tomcats state, it does not trigger a new check but seems to rely on “old” information from the cib; the tomcat resource script on the master remains untouched.

Any help welcome.

Thanks,

Andreas
PS: Some Version Information...
ubuntu 12.04 LTS
tomcat7 7.0.26-1ubuntu1 with self-written resource agent
pacemaker 1.1.6-2ubuntu3
corosync 1.4.2-2
resource-agents 1:3.9.2-5ubuntu4.1


--
CONET Solutions GmbH
Andreas Stallmann,
Theodor-Heuss-Allee 19, 53773 Hennef
Tel.: +49 2242 939-677, Fax: +49 2242 939-393
Mobil: +49 172 2455051
Internet: http://www.conet.de, mailto: AStallmann [at] CONET<mailto:AStallmann [at] CONET>



----------------------------
CONET Solutions GmbH, Theodor-Heuss-Allee 19, 53773 Hennef.
Registergericht/Registration Court: Amtsgericht Siegburg (HRB Nr. 9136)
Geschäftsführer/Managing Director: Anke Höfer

----------------------------


shyam.kaushik at gmail

Jul 2, 2012, 2:39 AM

Post #2 of 9 (407 views)
Permalink
Re: Resource is monitored only till promote... bug or config error? [In reply to]
I tested this patch & this solves the original bug of #5072. But I
also dont know what it means that the regression test suite is broken.

--Shyam

On Mon, Jul 2, 2012 at 2:53 PM, Stallmann, Andreas <AStallmann [at] conet> wrote:
> Hi!
>
>
>
> -----------------------------
> CONET Solutions GmbH, Theodor-Heuss-Allee 19, 53773 Hennef.
> Registergericht/Registration Court: Amtsgericht Siegburg (HRB Nr. 9136)
> Geschftsfhrer/Managing Directors: Anke Hfer
> -----------------------------
>
> -----Ursprngliche Nachricht-----
> Von: David Vossel [mailto:dvossel [at] redhat]
>> If you are patching an older version of pacemaker, this patch is probably all you need.
>> https://github.com/davidvossel/pacemaker/commit/fc03be02bf3a045babfe8233cbc99227da71d024
>
> Thanks, that was quick. :-) Still, Andrew Beekhof comments on this:
>
>> This breaks the regression tests.
>> Please make sure you run/update them before doing pull requests.
>
> What does this mean? Andrew, is this patch usable or not?
>
> Cheers,
>
> Andreas
> PS: Is any of the Ubuntu maintainers around in this mailing list? If yes: When will this patch make it to Ubuntu 12.04 LTS?
> _______________________________________________
> Pacemaker mailing list: Pacemaker [at] oss
> http://oss.clusterlabs.org/mailman/listinfo/pacemaker
>
> Project Home: http://www.clusterlabs.org
> Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
> Bugs: http://bugs.clusterlabs.org

_______________________________________________
Pacemaker mailing list: Pacemaker [at] oss
http://oss.clusterlabs.org/mailman/listinfo/pacemaker

Project Home: http://www.clusterlabs.org
Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
Bugs: http://bugs.clusterlabs.org


shyam.kaushik at gmail

Jul 2, 2012, 3:27 AM

Post #3 of 9 (406 views)
Permalink
Re: Resource is monitored only till promote... bug or config error? [In reply to]
Hi Andreas,

Its as good as patching any debian patches on ubuntu.

You could do something like this on your 12.04:
$ apt-get install devscripts
$ apt-get build-dep pacemaker
$ apt-get source pacemaker

now you will have the pacemaker sources
$ dch --local test (or whatever name you want labelled)
<apply the patch changes manually>
$ dpkg-source --commit (to get a patch)
$ debuild -us -uc

you should have all pacemaker packages & should be set at this step.

--Shyam



On Mon, Jul 2, 2012 at 3:42 PM, Stallmann, Andreas <AStallmann [at] conet> wrote:
> Hi Shyam,
>
>
>
> -----------------------------
> CONET Solutions GmbH, Theodor-Heuss-Allee 19, 53773 Hennef.
> Registergericht/Registration Court: Amtsgericht Siegburg (HRB Nr. 9136)
> Geschftsfhrer/Managing Directors: Anke Hfer
> -----------------------------
>
> -----Ursprngliche Nachricht-----
> Von: Shyam [mailto:shyam.kaushik [at] gmail]
>> I tested this patch & this solves the original bug of #5072. But I also dont know what it means that the regression test suite is broken.
>
> That's good to hear (well, at least the first bit... ;-) )
> I'm new to github. How can you download the patch only (and not the full source)?
>
> Cheers,
>
> Andreas
>
> _______________________________________________
> Pacemaker mailing list: Pacemaker [at] oss
> http://oss.clusterlabs.org/mailman/listinfo/pacemaker
>
> Project Home: http://www.clusterlabs.org
> Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
> Bugs: http://bugs.clusterlabs.org

_______________________________________________
Pacemaker mailing list: Pacemaker [at] oss
http://oss.clusterlabs.org/mailman/listinfo/pacemaker

Project Home: http://www.clusterlabs.org
Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
Bugs: http://bugs.clusterlabs.org


shyam.kaushik at gmail

Jul 2, 2012, 4:27 AM

Post #4 of 9 (408 views)
Permalink
Re: Resource is monitored only till promote... bug or config error? [In reply to]
Hi Andreas,

If you are not comfortable with git, the simplest would be to do "wget
https://raw.github.com/davidvossel/pacemaker/fc03be02bf3a045babfe8233cbc99227da71d024/lib/pengine/unpack.c"
& port the change in calculate_active_ops() manually.

--Shyam

On Mon, Jul 2, 2012 at 4:39 PM, Stallmann, Andreas <AStallmann [at] conet> wrote:
> Thanks, Shyam,
>
> that would have been my next question. But again to my FIRST one. :-) How do I download the patch? There's no download link anywhere. Do I simply have to copy-and-paste from the text box?
>
> Cheers,
>
> Andreas
> --
> CONET Solutions GmbH
> Andreas Stallmann,
> Theodor-Heuss-Allee 19, 53773 Hennef
> Tel.: +49 2242 939-677, Fax: +49 2242 939-393
> Mobil: +49 172 2455051
> Internet: http://www.conet.de, mailto: AStallmann [at] CONET
>
> -----Ursprngliche Nachricht-----
> Von: Shyam [mailto:shyam.kaushik [at] gmail]
> Gesendet: Montag, 2. Juli 2012 12:27
> An: The Pacemaker cluster resource manager
> Betreff: Re: [Pacemaker] Resource is monitored only till promote... bug or config error?
>
> Hi Andreas,
>
> Its as good as patching any debian patches on ubuntu.
>
> You could do something like this on your 12.04:
> $ apt-get install devscripts
> $ apt-get build-dep pacemaker
> $ apt-get source pacemaker
>
> now you will have the pacemaker sources
> $ dch --local test (or whatever name you want labelled) <apply the patch changes manually> $ dpkg-source --commit (to get a patch) $ debuild -us -uc
>
> you should have all pacemaker packages & should be set at this step.
>
> --Shyam
>
>
>
> On Mon, Jul 2, 2012 at 3:42 PM, Stallmann, Andreas <AStallmann [at] conet> wrote:
>> Hi Shyam,
>>
>>
>>
>> -----------------------------
>> CONET Solutions GmbH, Theodor-Heuss-Allee 19, 53773 Hennef.
>> Registergericht/Registration Court: Amtsgericht Siegburg (HRB Nr.
>> 9136) Geschftsfhrer/Managing Directors: Anke Hfer
>> -----------------------------
>>
>> -----Ursprngliche Nachricht-----
>> Von: Shyam [mailto:shyam.kaushik [at] gmail]
>>> I tested this patch & this solves the original bug of #5072. But I also dont know what it means that the regression test suite is broken.
>>
>> That's good to hear (well, at least the first bit... ;-) ) I'm new to
>> github. How can you download the patch only (and not the full source)?
>>
>> Cheers,
>>
>> Andreas
>>
>> _______________________________________________
>> Pacemaker mailing list: Pacemaker [at] oss
>> http://oss.clusterlabs.org/mailman/listinfo/pacemaker
>>
>> Project Home: http://www.clusterlabs.org Getting started:
>> http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
>> Bugs: http://bugs.clusterlabs.org
>
> _______________________________________________
> Pacemaker mailing list: Pacemaker [at] oss http://oss.clusterlabs.org/mailman/listinfo/pacemaker
>
> Project Home: http://www.clusterlabs.org Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
> Bugs: http://bugs.clusterlabs.org
>
> _______________________________________________
> Pacemaker mailing list: Pacemaker [at] oss
> http://oss.clusterlabs.org/mailman/listinfo/pacemaker
>
> Project Home: http://www.clusterlabs.org
> Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
> Bugs: http://bugs.clusterlabs.org

_______________________________________________
Pacemaker mailing list: Pacemaker [at] oss
http://oss.clusterlabs.org/mailman/listinfo/pacemaker

Project Home: http://www.clusterlabs.org
Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
Bugs: http://bugs.clusterlabs.org


andrew at beekhof

Jul 2, 2012, 4:44 AM

Post #5 of 9 (403 views)
Permalink
Re: Resource is monitored only till promote... bug or config error? [In reply to]
On Mon, Jul 2, 2012 at 7:23 PM, Stallmann, Andreas <AStallmann [at] conet> wrote:
> Hi!
>
>
>
> -----------------------------
> CONET Solutions GmbH, Theodor-Heuss-Allee 19, 53773 Hennef.
> Registergericht/Registration Court: Amtsgericht Siegburg (HRB Nr. 9136)
> Geschftsfhrer/Managing Directors: Anke Hfer
> -----------------------------
>
> -----Ursprngliche Nachricht-----
> Von: David Vossel [mailto:dvossel [at] redhat]
>> If you are patching an older version of pacemaker, this patch is probably all you need.
>> https://github.com/davidvossel/pacemaker/commit/fc03be02bf3a045babfe8233cbc99227da71d024
>
> Thanks, that was quick. :-) Still, Andrew Beekhof comments on this:
>
>> This breaks the regression tests.
>> Please make sure you run/update them before doing pull requests.
>
> What does this mean? Andrew, is this patch usable or not?

It fixes the bug, but /may/ (re-)introduce another.
David needs to look into the test results and verify that they are correct.

>
> Cheers,
>
> Andreas
> PS: Is any of the Ubuntu maintainers around in this mailing list? If yes: When will this patch make it to Ubuntu 12.04 LTS?
> _______________________________________________
> Pacemaker mailing list: Pacemaker [at] oss
> http://oss.clusterlabs.org/mailman/listinfo/pacemaker
>
> Project Home: http://www.clusterlabs.org
> Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
> Bugs: http://bugs.clusterlabs.org

_______________________________________________
Pacemaker mailing list: Pacemaker [at] oss
http://oss.clusterlabs.org/mailman/listinfo/pacemaker

Project Home: http://www.clusterlabs.org
Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
Bugs: http://bugs.clusterlabs.org


dvossel at redhat

Jul 2, 2012, 7:14 AM

Post #6 of 9 (406 views)
Permalink
Re: Resource is monitored only till promote... bug or config error? [In reply to]
----- Original Message -----
> From: "Shyam" <shyam.kaushik [at] gmail>
> To: "The Pacemaker cluster resource manager" <pacemaker [at] oss>
> Sent: Monday, July 2, 2012 4:39:54 AM
> Subject: Re: [Pacemaker] Resource is monitored only till promote... bug or config error?
>
> I tested this patch & this solves the original bug of #5072. But I
> also dont know what it means that the regression test suite is
> broken.
>
> --Shyam

Shyam,

Thanks for testing the patch. Unfortunately the regression test failures are indicating that there is more work to be done. I would not recommend using this patch in production until the regression test failures have been sorted out. I will be investigating this further today.

-- Vossel

_______________________________________________
Pacemaker mailing list: Pacemaker [at] oss
http://oss.clusterlabs.org/mailman/listinfo/pacemaker

Project Home: http://www.clusterlabs.org
Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
Bugs: http://bugs.clusterlabs.org


shyam.kaushik at gmail

Jul 2, 2012, 8:49 AM

Post #7 of 9 (404 views)
Permalink
Re: Resource is monitored only till promote... bug or config error? [In reply to]
Hi Andreas, for sure this issue never happened with pacemaker
1.0.10-4ubuntu3 available in Ubuntu Natty. This started happening once
we moved to precise's version.

Hi David, let me know if you need me to run some debug enabled change
of yours or so. Thanks.

--Shyam

On Mon, Jul 2, 2012 at 8:44 PM, David Vossel <dvossel [at] redhat> wrote:
> ----- Original Message -----
>> From: "Andreas Stallmann" <AStallmann [at] CONET>
>> To: "The Pacemaker cluster resource manager" <pacemaker [at] oss>
>> Sent: Monday, July 2, 2012 9:54:57 AM
>> Subject: Re: [Pacemaker] Resource is monitored only till promote... bug or config error?
>>
>> Hi David,
>>
>> do you by chance know, if this bug is not there in earlier versions?
>> If yes, we would like to downgrade our system to this version.
>>
>> Any other ideas for a workaround?
>
> I am not aware of a version this works in, or a good workaround. This is a high priority for me at the moment and I have a good understanding of why it occurs. I expect to have a new patch out in the next day or two. I'd recommend wanting for that.
>
> -- Vossel
>
>> Thanks,
>>
>> Andreas
>>
>>
>> -----------------------------
>> CONET Solutions GmbH, Theodor-Heuss-Allee 19, 53773 Hennef.
>> Registergericht/Registration Court: Amtsgericht Siegburg (HRB Nr.
>> 9136)
>> Gesch?ftsf?hrer/Managing Directors: Anke H?fer
>> -----------------------------
>>
>>
>> _______________________________________________
>> Pacemaker mailing list: Pacemaker [at] oss
>> http://oss.clusterlabs.org/mailman/listinfo/pacemaker
>>
>> Project Home: http://www.clusterlabs.org
>> Getting started:
>> http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
>> Bugs: http://bugs.clusterlabs.org
>>
>
> _______________________________________________
> Pacemaker mailing list: Pacemaker [at] oss
> http://oss.clusterlabs.org/mailman/listinfo/pacemaker
>
> Project Home: http://www.clusterlabs.org
> Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
> Bugs: http://bugs.clusterlabs.org

_______________________________________________
Pacemaker mailing list: Pacemaker [at] oss
http://oss.clusterlabs.org/mailman/listinfo/pacemaker

Project Home: http://www.clusterlabs.org
Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
Bugs: http://bugs.clusterlabs.org


dvossel at redhat

Jul 2, 2012, 1:53 PM

Post #8 of 9 (403 views)
Permalink
Re: Resource is monitored only till promote... bug or config error? [In reply to]
----- Original Message -----
> From: "Shyam" <shyam.kaushik [at] gmail>
> To: "The Pacemaker cluster resource manager" <pacemaker [at] oss>
> Sent: Monday, July 2, 2012 10:49:50 AM
> Subject: Re: [Pacemaker] Resource is monitored only till promote... bug or config error?
>
> Hi Andreas, for sure this issue never happened with pacemaker
> 1.0.10-4ubuntu3 available in Ubuntu Natty. This started happening
> once
> we moved to precise's version.
>
> Hi David, let me know if you need me to run some debug enabled change
> of yours or so. Thanks.
>
> --Shyam


Okay, I believe I got this sorted out. Use the updated patch below if you are interested in testing.

https://github.com/davidvossel/pacemaker/commit/5d9344042ac3727badf5e26c29aff0fa6b71fcc8

If you run the pengine regression tests, one test will fail still. I investigated the cause of the failure and it appears to be a case where the test results just need to be updated to reflect the new changes.

-- Vossel



_______________________________________________
Pacemaker mailing list: Pacemaker [at] oss
http://oss.clusterlabs.org/mailman/listinfo/pacemaker

Project Home: http://www.clusterlabs.org
Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
Bugs: http://bugs.clusterlabs.org


shyam.kaushik at gmail

Jul 2, 2012, 10:38 PM

Post #9 of 9 (406 views)
Permalink
Re: Resource is monitored only till promote... bug or config error? [In reply to]
Thanks a lot David.

I confirmed that this fix works & observed that regression test fails
on one test case like you mentioned. I will use this fix.

--Shyam

On Tue, Jul 3, 2012 at 2:23 AM, David Vossel <dvossel [at] redhat> wrote:
> ----- Original Message -----
>> From: "Shyam" <shyam.kaushik [at] gmail>
>> To: "The Pacemaker cluster resource manager" <pacemaker [at] oss>
>> Sent: Monday, July 2, 2012 10:49:50 AM
>> Subject: Re: [Pacemaker] Resource is monitored only till promote... bug or config error?
>>
>> Hi Andreas, for sure this issue never happened with pacemaker
>> 1.0.10-4ubuntu3 available in Ubuntu Natty. This started happening
>> once
>> we moved to precise's version.
>>
>> Hi David, let me know if you need me to run some debug enabled change
>> of yours or so. Thanks.
>>
>> --Shyam
>
>
> Okay, I believe I got this sorted out. Use the updated patch below if you are interested in testing.
>
> https://github.com/davidvossel/pacemaker/commit/5d9344042ac3727badf5e26c29aff0fa6b71fcc8
>
> If you run the pengine regression tests, one test will fail still. I investigated the cause of the failure and it appears to be a case where the test results just need to be updated to reflect the new changes.
>
> -- Vossel
>
>
>
> _______________________________________________
> Pacemaker mailing list: Pacemaker [at] oss
> http://oss.clusterlabs.org/mailman/listinfo/pacemaker
>
> Project Home: http://www.clusterlabs.org
> Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
> Bugs: http://bugs.clusterlabs.org

_______________________________________________
Pacemaker mailing list: Pacemaker [at] oss
http://oss.clusterlabs.org/mailman/listinfo/pacemaker

Project Home: http://www.clusterlabs.org
Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
Bugs: http://bugs.clusterlabs.org

Linux-HA pacemaker RSS feed   Index | Next | Previous | View Threaded
 
 


Interested in having your list archived? Contact Gossamer Threads
 
  Web Applications & Managed Hosting Powered by Gossamer Threads Inc.