[Pacemaker] Some errors after upgrading from heartbeat 2.1 cluster

Drew Morone andrewmorone at yahoo.com
Tue Jul 24 13:59:09 UTC 2012


I replaced "target_role" with "target-role", but I get the same error regarding "target-role" (eg. ERROR: DRBD_data: attribute target_role does not exist).
I also tried removing that attribute, but it seems to break my config. Should I post the whole XML? I don't want to spam the list if it's not necessary.


Drew



________________________________

Date: Mon, 23 Jul 2012 14:17:08 -0400 (EDT)
From: David Vossel <dvossel at redhat.com>
To: The Pacemaker cluster resource manager
    <pacemaker at oss.clusterlabs.org>
Subject: Re: [Pacemaker] Some errors after upgrading from heartbeat
    2.1 cluster
Message-ID: <1268593649.150738.1343067428434.JavaMail.root at redhat.com>
Content-Type: text/plain; charset=utf-8

----- Original Message -----
> From: "Drew Morone" <andrewmorone at yahoo.com>
> To: pacemaker at oss.clusterlabs.org
> Sent: Monday, July 23, 2012 10:51:23 AM
> Subject: [Pacemaker] Some errors after upgrading from heartbeat 2.1 cluster
> 
> 
> 
> 
> Hi,
> I'm new to this list and have been searching for a solution for a few
> weeks to a problem I'm having, but not seeing anything similar to
> what i'm seeing.
> 
> 
> We have a 2 node active/standby cluster with the following services:
> drbd
> nfs
> dhcpd
> 
> 
> We currently upgraded our OS from Debian Lenny, to Squeeze, which in
> turn upgraded heartbeat to hearbeat(3.0.3) /pacemaker(1.0.9) . The
> upgrade actually went pretty well following this example:
> http://www.linux-ha.org/doc/users-guide/_upgrading_from_crm_enabled_heartbeat_2_1_clusters.html
> 
> and everything seems to work well, no down time, no loss of data. My
> only issue is that I'm getting some errors with some of the resource
> attributes and parameters, eg running 'crm configure verify' I'll
> get stuff like this:
> 
> ERROR: DRBD_data: attribute target_role does not exist
> ERROR: DRBD_data: parameter target_role does not exist
> 
> 
> 
> Here's the section of the converted schema it's complaining about:
> <primitive id="DRBD_data" class="heartbeat" type="drbddisk">
> <meta_attributes id="primitive-DRBD_data.meta">
> <nvpair id="nvpair.meta.auto-19" name="target_role" value="started"/>

I believe the meta attribute is target-role not target_role.

-- Vossel

> </meta_attributes>
> <instance_attributes id="instance_attributes.id179586">
> <nvpair id="nvpair.id178988" name="target_role" value="started"/>
> <nvpair name="target_role" value="drbd0" id="nvpair.id178172"/>
> </instance_attributes>
> </primitive>
> 
> 
> I'm assuming that those parameters are no longer used, or used
> differently in the heartbeat/pacemaker configurations. We upgraded
> to the new schema format. I don't want to rebuild the configuration
> from scratch, but I'm thinking maybe I might have to. Unless these
> errors are benign. Everything seems to work properly.
> 
> 
> Drew
> _______________________________________________
> Pacemaker mailing list: Pacemaker at oss.clusterlabs.org
> 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
> 



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

Message: 2
Date: Mon, 23 Jul 2012 12:38:21 -0700
From: David Barchas <dave at barchas.com>
To: pacemaker at oss.clusterlabs.org
Subject: Re: [Pacemaker] None of the standard agents in ocf:heartbeat
    are working in centos 6
Message-ID: <2EDD049E45D549E69571B95D5F355D72 at barchas.com>
Content-Type: text/plain; charset="utf-8"



On Monday, July 23, 2012 at 7:48 AM, David Barchas wrote:

> 
> > Date: Mon, 23 Jul 2012 14:15:27 +0300
> > From: Vladislav Bogdanov
> > 
> > 23.07.2012 08:06, David Barchas wrote:
> > > Hello.
> > > 
> > > I have been working on this for 3 days now, and must be so stressed out
> > > that I am being blinded to what is probably an obvious cause of this. In
> > > a word, HELP.
> > > 
> > 
> > 
> > setenforce 0 ?
> > 
> i am familiar with it but have never had to disable it. I would be surprised for packages in standard repos.
> However, I am going to try that anyway just to make sure. 
> 
> thanks for the help!
> dave
> 

That worked...
Aside from the fact that I have probably spent about 5 hours per each letter of that solution and that nowhere on the tubes is that even mentioned regarding the base agents, I am grateful. Thank you. You may have just saved me from having a stroke. 

Now however, I need to understand why?
What are my risks? I imagine none, since this is storage server with almost no access.
What does this tell me was actually wrong? Do I blame SElinux or the agents or pacemaker? Any way to figure it out?

thanks again, and the rest of you who lent your brains to me.

dave



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://oss.clusterlabs.org/pipermail/pacemaker/attachments/20120723/9e5d1a47/attachment-0001.html>

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

Message: 3
Date: Mon, 23 Jul 2012 21:42:10 +0200
From: Marcus Osdoba <marcus.osdoba at googlemail.com>
To: The Pacemaker cluster resource manager
    <pacemaker at oss.clusterlabs.org>
Subject: [Pacemaker] lxc resource fixes
Message-ID: <500DA912.3010209 at googlemail.com>
Content-Type: text/plain; charset=ISO-8859-15; format=flowed

Dear mailinglist,

I am using lxc in a pacemaker clustered environment.
I have noted, that the current version in git is not capable to work 
with 0.7.5 correctly.

The lxc-ps command changed in 0.7.5, too. So I have applied the same fix 
as the last commit on lxc in resource-agent.git [0].

[1] and [2] are cosmetical fixes. I think it's better to give a warning, 
when the container is stopped instead of cleanly halted.

Regarding [3], at least in Debian based systems, it's possible to launch 
the lxc-halt command to perform a clean shutdown of the container. 
Therfore, I enhanced the current lxc-ocf version to support an additonal 
optional parameter named shutdownmethod. It defaults to the current 
behaviour, which is sending SIGPWR to init/systemd processes inside the 
container. Alternatively, it's possible to command an lxc-halt to the 
container (which does not require the requested powerfail-line in inittab).


Please consider committing [0] at least and think about the alternative 
halt-method [3].

Thanks in advance and kind regards,
Marcus


[0] 
https://gitorious.org/debian/resource-agents/commit/f7b8fba8809d34c537e95a1aeaa2bc64463a9049

[1] 
https://gitorious.org/debian/resource-agents/commit/08c321a5c92793d6b43bbffef980fdabaa1df1ce

[2] 
https://gitorious.org/debian/resource-agents/commit/42838ca854557a06ee9669cea9e1c560b5b50d46

[3] 
https://gitorious.org/debian/resource-agents/commit/cc3bdb5968d0a28a695acc93c23a4cf1f09749f6



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

Message: 4
Date: Tue, 24 Jul 2012 11:50:54 +1000
From: Andrew Beekhof <andrew at beekhof.net>
To: The Pacemaker cluster resource manager
    <pacemaker at oss.clusterlabs.org>
Subject: Re: [Pacemaker] None of the standard agents in ocf:heartbeat
    are working in centos 6
Message-ID:
    <CAEDLWG1L8_6BbtzP=8QoQiMFt3Pp8BcWGnEKD+T_0xUUha5jMg at mail.gmail.com>
Content-Type: text/plain; charset=ISO-8859-1

On Tue, Jul 24, 2012 at 5:38 AM, David Barchas <dave at barchas.com> wrote:
>
> On Monday, July 23, 2012 at 7:48 AM, David Barchas wrote:
>
>
> Date: Mon, 23 Jul 2012 14:15:27 +0300
> From: Vladislav Bogdanov
>
> 23.07.2012 08:06, David Barchas wrote:
>
> Hello.
>
> I have been working on this for 3 days now, and must be so stressed out
> that I am being blinded to what is probably an obvious cause of this. In
> a word, HELP.
>
>
> setenforce 0 ?
>
> i am familiar with it but have never had to disable it. I would be surprised
> for packages in standard repos.

No-one has written an selinux policy for pacemaker yet.
I would imagine that will come in the next month or so.

> However, I am going to try that anyway just to make sure.
>
> thanks for the help!
> dave
>
>
> That worked...
> Aside from the fact that I have probably spent about 5 hours per each letter
> of that solution and that nowhere on the tubes is that even mentioned
> regarding the base agents, I am grateful. Thank you. You may have just saved
> me from having a stroke.
>
> Now however, I need to understand why?
> What are my risks? I imagine none, since this is storage server with almost
> no access.
> What does this tell me was actually wrong? Do I blame SElinux or the agents
> or pacemaker? Any way to figure it out?
>
> thanks again, and the rest of you who lent your brains to me.
>
> dave
>
>
>
>
> _______________________________________________
> Pacemaker mailing list: Pacemaker at oss.clusterlabs.org
> 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
>



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

Message: 5
Date: Mon, 23 Jul 2012 19:32:03 -0700
From: David Barchas <dave at barchas.com>
To: The Pacemaker cluster resource manager
    <pacemaker at oss.clusterlabs.org>
Subject: Re: [Pacemaker] None of the standard agents in ocf:heartbeat
    are working in centos 6
Message-ID: <-4283991851919095708 at unknownmsgid>
Content-Type: text/plain; charset=ISO-8859-1

--Sent mobile--

On Jul 23, 2012, at 6:58 PM, Andrew Beekhof <andrew at .net> wrote:

> On Tue, Jul 24, 2012 at 5:38 AM, David Barchas <dave at .com> wrote:
>>
>> On Monday, July 23, 2012 at 7:48 AM, David Barchas wrote:
>>
>>
>> Date: Mon, 23 Jul 2012 14:15:27 +0300
>> From: Vladislav Bogdanov
>>
>> 23.07.2012 08:06, David Barchas wrote:
>>
>> Hello.
>>
>> I have been working on this for 3 days now, and must be so stressed out
>> that I am being blinded to what is probably an obvious cause of this. In
>> a word, HELP.
>>
>>
>> setenforce 0 ?
>>
>> i am familiar with it but have never had to disable it. I would be surprised
>> for packages in standard repos.
>
> No-one has written an selinux policy for pacemaker yet.
> I would imagine that will come in the next month or so.

Ahhhh. I have never had to deal with SELinux before. Other
coro/pace/drbd installs of mine have been other less enterprise level
distros like Ubuntu. Simple things always get me.

>
>> However, I am going to try that anyway just to make sure.
>>
>> thanks for the help!
>> dave
>>
>>
>> That worked...
>> Aside from the fact that I have probably spent about 5 hours per each letter
>> of that solution and that nowhere on the tubes is that even mentioned
>> regarding the base agents, I am grateful. Thank you. You may have just saved
>> me from having a stroke.
>>
>> Now however, I need to understand why?
>> What are my risks? I imagine none, since this is storage server with almost
>> no access.
>> What does this tell me was actually wrong? Do I blame SElinux or the agents
>> or pacemaker? Any way to figure it out?
>>
>> thanks again, and the rest of you who lent your brains to me.
>>
>> dave
>>
>>
>>
>>
>> _______________________________________________
>> Pacemaker mailing list: Pacemaker at oss.clusterlabs.org
>> 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.clusterlabs.org
> 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



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

Message: 6
Date: Tue, 24 Jul 2012 13:10:43 +0900
From: Yuichi SEINO <seino.cluster2 at gmail.com>
To: Jiaju Zhang <jjzhang at suse.de>
Cc: The Pacemaker cluster resource manager
    <pacemaker at oss.clusterlabs.org>
Subject: Re: [Pacemaker] Do you need to execute crm_ticket on
    arbitrator?
Message-ID:
    <CAMb0o5Lk+Yj8=5rX3AxWHrk7DgvaHuTeQYngho3Mm-sex26ejA at mail.gmail.com>
Content-Type: text/plain; charset=UTF-8

> Yes, arbitrator is not using cib. I'm fine with removing the crm_ticket
> calling logic on arbitrator. An implementation detail on this issue
> would be to keep the modularity of the code structure, that is,
> proposer, acceptor and learner is the terminology of paxos algorithm,
> site and arbitrator is the GEO clustering terminology, I'd like each
> component to be more independent and its internal logic is transparent
> to other components.
>
> Thanks,
> Jiaju
>
>

Hi jiaju,

I understand the answer. Do you fix this problem on the next development?

And, I know that the following function will be developed when you
work the next development.

* new catch-up code
* the refactoring of the transport layer

I would like to know if there are other functions.

Sincerely,
Yuichi

-- 
Yuichi SEINO
METROSYSTEMS CORPORATION
E-mail:seino.cluster2 at gmail.com



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

Message: 7
Date: Tue, 24 Jul 2012 12:25:05 +0800
From: Jiaju Zhang <jjzhang at suse.de>
To: Yuichi SEINO <seino.cluster2 at gmail.com>
Cc: The Pacemaker cluster resource manager
    <pacemaker at oss.clusterlabs.org>
Subject: Re: [Pacemaker] Do you need to execute crm_ticket on
    arbitrator?
Message-ID: <1343103905.3396.5.camel at linux-jjzhang.site>
Content-Type: text/plain; charset="UTF-8"

On Tue, 2012-07-24 at 13:10 +0900, Yuichi SEINO wrote:
> > Yes, arbitrator is not using cib. I'm fine with removing the crm_ticket
> > calling logic on arbitrator. An implementation detail on this issue
> > would be to keep the modularity of the code structure, that is,
> > proposer, acceptor and learner is the terminology of paxos algorithm,
> > site and arbitrator is the GEO clustering terminology, I'd like each
> > component to be more independent and its internal logic is transparent
> > to other components.
> >
> > Thanks,
> > Jiaju
> >
> >
> 
> Hi jiaju,
> 
> I understand the answer. Do you fix this problem on the next development?

Yes.

> 
> And, I know that the following function will be developed when you
> work the next development.
> 
> * new catch-up code
> * the refactoring of the transport layer
> 
> I would like to know if there are other functions.

To add more policies and make booth more configurable, add watchdog
daemon.

Also you can raise your feature request if you think something need to
be implemented;)

Thanks,
Jiaju





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

Message: 8
Date: Tue, 24 Jul 2012 07:25:24 +0300
From: Vladislav Bogdanov <bubble at hoster-ok.com>
To: The Pacemaker cluster resource manager
    <pacemaker at oss.clusterlabs.org>
Subject: Re: [Pacemaker] None of the standard agents in ocf:heartbeat
    are working in centos 6
Message-ID: <500E23B4.8070901 at hoster-ok.com>
Content-Type: text/plain; charset=ISO-8859-1

24.07.2012 04:50, Andrew Beekhof wrote:
> On Tue, Jul 24, 2012 at 5:38 AM, David Barchas <dave at barchas.com> wrote:
>>
>> On Monday, July 23, 2012 at 7:48 AM, David Barchas wrote:
>>
>>
>> Date: Mon, 23 Jul 2012 14:15:27 +0300
>> From: Vladislav Bogdanov
>>
>> 23.07.2012 08:06, David Barchas wrote:
>>
>> Hello.
>>
>> I have been working on this for 3 days now, and must be so stressed out
>> that I am being blinded to what is probably an obvious cause of this. In
>> a word, HELP.
>>
>>
>> setenforce 0 ?
>>
>> i am familiar with it but have never had to disable it. I would be surprised
>> for packages in standard repos.
> 
> No-one has written an selinux policy for pacemaker yet.
> I would imagine that will come in the next month or so.
> 

Highly appreciated. However lrmd part may be not as easy to implement
properly as it seems at the first glance.





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

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


End of Pacemaker Digest, Vol 56, Issue 54
*****************************************
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.clusterlabs.org/pipermail/pacemaker/attachments/20120724/242da4ab/attachment.htm>


More information about the Pacemaker mailing list