gmane.linux.highavailability.user

Re: Resource Collocation v/s Resource Groups

Andrew Beekhof — 2013-06-19T22:10:57

On 18/06/2013, at 7:18 PM, Parkirat <parkiratbagga< at >gmail.com> wrote:


Yes they are.  There are 7 of them:

meta target-role="Started" \
meta is-managed="true" target-role="Started" \
meta resource-stickiness="100" target-role="Started" \
meta migration-threshold="2" target-role="Started" \
meta multiple-active="stop_start" target-role="Started" \
meta failure-timeout="300" target-role="Started" \
meta allow-migrate="true" target-role="Started"


no


dont


Monitors don't start anything.  They detect the failure and prompt pacemaker to initiate recovery.
Please attach logs if you think pacemaker is not recovering the apache process correctly.


_______________________________________________
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

Ipaddr2 and VLAN interfaces ?

Jakob Curdes — 2013-06-19T14:40:47

Hi, I amtrying to setup  IPaddr2 resources on a cman-managed cluster 
which need to be bound to VLAN interfaces on linux. I have something like

bond1 with a static address 192.168.10.2
bond1.0001 (unnumbered)

and want to configure a IPaddr2 resource  on bond1.0001.
However regardless what I try it does not work; I tried using the nic 
parameter and the label parameter, passing the VLAN ID as label:

pcs resource create VLAN_003_IP  ocf:heartbeat:IPaddr2 ip=192.168.11.2 
cidr_netmask=24 op monitor interval=30s nic=bond1 iflabel=0003
ERROR: Unable to create resource/fence device
Call cib_create failed (-76): Name not unique on network

It seems that in the old heartbeat v1 times it was possible to specify a 
VLAN interface with something like

192.168.1.1/29/eth0.30

How can I do this nowadays?

Thank you for hints,
Jakob Curdes

_______________________________________________
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: LVM Resource agent, "exclusive" activation

David Vossel — 2013-06-18T19:03:13




----- Original Message -----

I see stripping the tag on stop as cleaning up the magic the LVM agent is doing behind the scenes.  The tag is only needed to allow activation, activation is still prevented outside of the cluster management with or without the tag being present on the vg.


Yep, we could do this.  It puts another step in the admin's hands that we can automate though.


I gave this a lot of thought.  single-lv activation adds additional complexity to this whole exclusive activation with tags. Rather than complicate the situation any further, I am in favor of not pulling support for lv activation to the heartbeat agent.  This agent is already complex to manage as it is.

Re: LVM Resource agent, "exclusive" activation

Lars Ellenberg — 2013-06-18T09:30:30

Manny thanks for all the work,
as "tomorrow" has passed,
I just merged it.

Added a "tag" parameter description.

I'm not sure why we would need to "strip" the tag on stop though?

Or why we would override a different tag on start.

As the "cluster tag" is not supposed to change,
we could just require the admin to set it once.

Has the side-effect that an admin can revoke the "cluster rights"
by simply re-tagging with something != the cluster tag.

Do we still want the single-lv activation feature as well?

Re: Resource Collocation v/s Resource Groups

Parkirat — 2013-06-18T09:18:58

Hi Sven,

Thanks for replying back. 

By manually stopping, I mean, I am stopping the resource by running the
below command:
/etc/init.d/httpd stop, which is outside the control of pacemaker.

Also, the entries for target-role is not duplicate. Its configure once for
each resource. Moreover, I am not using group in the configuration but I am
just collocating the resources. In the configurations see :
colocation eip_for_apache inf: ElasticIP ApacheServer

I am now confused at the below points:

1. To use collocations do we also need to make the groups?
2. I tried groups and collocation separately but not combined. How to use
them in combination?
3. With both using groups and collocation, if 1 resource is externally
(outside the control of pacemaker) stopped, why is it not getting re-started
by the pacemaker, when there is a monitor configured for resource. But if I
do not use either of two viz. groups & collocation and just configure the
resource not associating with any other resource, it does get started by the
monitor properly.

Regards,
Parkirat Singh Bagga.



--
View this message in context: http://linux-ha.996297.n3.nabble.com/Resource-Collocation-v-s-Resource-Groups-tp14699p14707.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

Re: Resource Collocation v/s Resource Groups

Sven Arnold — 2013-06-18T08:38:09

Disclaimer: I am also new and have not used migration-threshold yet.

But I am not sure what you mean by "manually stopped". Do you mean by 
issuing a manual command "resource stop" at the crm shell? If so, the 
resource will remain stopped until you issue a new start command.

If you stop the resource outside the control of pacemaker (e.g. by 
killing the service) I would also expect that the resource manager would 
try to restart it after monitor interval + timeout seconds.

Looking at your config I am a bit confused that there are so many lines 
starting with "meta". Also there are duplicated entries of 
target-role="Started". I would suggest to collect all meta attributes 
after _one_ meta keyword per resource.

Additionally, at least when editing the configuration to move resources 
into groups you should remove remaining "target-role" entries from the 
resources themselves.

Regards,

Sven

_______________________________________________
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: LVM Resource agent, "exclusive" activation

David Vossel — 2013-06-17T21:53:57




----- Original Message -----

This patch is going in tomorrow. Speak up now if you have any reservations concerning it.

Re: Resource Collocation v/s Resource Groups

Parkirat — 2013-06-17T21:53:33

My configuration looks like as below:

[root< at >prod-hb-nmn-002 ~]# crm configure show
node $id="091a7142-4272-402a-aeff-ce1b25123668" prod-hb-nmn-002
node $id="7910c4de-718d-45d7-b4da-24b3b65b9855" prod-hb-nmn-001
primitive ApacheServer ocf:heartbeat:apache \
params configfile="/etc/httpd/conf/httpd.conf" \
op monitor interval="60s" timeout="120s" on-fail="restart" \
op start interval="0" timeout="40s" \
op stop interval="0" timeout="60s" \
meta target-role="Started" \
meta is-managed="true" \
meta resource-stickiness="100" \
meta migration-threshold="2" \
meta multiple-active="stop_start" \
meta failure-timeout="300" \
meta allow-migrate="true"
primitive ElasticIP ocf:heartbeat:elasticip \
op monitor interval="300s" timeout="120s" on-fail="restart" \
op start interval="0" timeout="60s" \
op stop interval="0" timeout="60s" \
meta target-role="Started" \
meta is-managed="true" target-role="Started" \
meta resource-stickiness="100" target-role="Started" \
meta migration-threshold="2" target-role="Started" \
meta multiple-active="stop_start" target-role="Started" \
meta failure-timeout="300" target-role="Started" \
meta allow-migrate="true" target-role="Started"
colocation eip_for_apache inf: ElasticIP ApacheServer
property $id="cib-bootstrap-options" \
dc-version="1.0.9-89bd754939df5150de7cd76835f98fe90851b677" \
cluster-infrastructure="Heartbeat" \
no-quorum-policy="ignore" \
stonith-enabled="false" \
last-lrm-refresh="1371505017" \
default-resource-stickiness="100" \
expected-quorum-votes="2"
rsc_defaults $id="rsc-options" \
resource-stickiness="100"



--
View this message in context: http://linux-ha.996297.n3.nabble.com/Resource-Collocation-v-s-Resource-Groups-tp14699p14704.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

Re: Resource Collocation v/s Resource Groups

Parkirat — 2013-06-17T21:49:12

Hi Sven,

Thanks for the reply. I am now collocating resources but the same problem of
the resource not getting started when manually stopped persists when my
migration_threshold=2, and its the 1st time, I have brought down the
resource after doing the cleanup as well as waiting for the failure-timeout
on that node.

Note: It behaves properly when my resources are not collocated or grouped.

Regards,
Parkirat Singh Bagga.



--
View this message in context: http://linux-ha.996297.n3.nabble.com/Resource-Collocation-v-s-Resource-Groups-tp14699p14703.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

Re: Resource Collocation v/s Resource Groups

Sven Arnold — 2013-06-17T18:18:59

Hi Pakirat,


Resources inside a resource group are colocated _and_ ordered. See 
http://clusterlabs.org/doc/Ordering_Explained.pdf for more details.


Best regards,

Sven
_______________________________________________
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: Why cman is started at rc0.d and rc6.d

Su Chen — 2013-06-17T18:19:03

And another thing is it never get killed. Should it be stopped when system is halting?

root< at >SuTH3:/etc# ls /etc/rc* | grep cman
S05cman
S05cman
S61cman

Thanks,
Su

From: Su Chen
Sent: Monday, June 17, 2013 11:17 AM
To: 'General Linux-HA mailing list'
Subject: Why cman is started at rc0.d and rc6.d

Hi All,

I am very new to pacemaker, cororsync and cman. I installed the packages on an Ubuntu machine. ("aptitude install pacemaker cman fence-agents")
To my surprise, cman has a link under rc0.d and rc6.d, why cman need to be started while system is shuting down?

root< at >SuTH3:/etc# ls -l /etc/rc0.d/S05cman /etc/rc6.d/S05cman
lrwxrwxrwx 1 root root 14 May 18 23:55 /etc/rc0.d/S05cman -> ../init.d/cman
lrwxrwxrwx 1 root root 14 May 18 23:55 /etc/rc6.d/S05cman -> ../init.d/cman

Thanks,
Su
_______________________________________________
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

Why cman is started at rc0.d and rc6.d

Su Chen — 2013-06-17T18:16:31

Hi All,

I am very new to pacemaker, cororsync and cman. I installed the packages on an Ubuntu machine. ("aptitude install pacemaker cman fence-agents")
To my surprise, cman has a link under rc0.d and rc6.d, why cman need to be started while system is shuting down?

root< at >SuTH3:/etc# ls -l /etc/rc0.d/S05cman /etc/rc6.d/S05cman
lrwxrwxrwx 1 root root 14 May 18 23:55 /etc/rc0.d/S05cman -> ../init.d/cman
lrwxrwxrwx 1 root root 14 May 18 23:55 /etc/rc6.d/S05cman -> ../init.d/cman

Thanks,
Su
_______________________________________________
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

Resource Collocation v/s Resource Groups

Parkirat — 2013-06-17T18:12:24

Hi,

Is there any difference between Resource Collocation and Resource Groups?

I grouped 2 resources both having migration_threshold=2 and
monitor_interval=60s. When, I stopped 1 of the resource from the group, I
did not restarted. However, when I was configuring the resource not in the
group, the resource started on manually stopping it.

Also is there any way to order the sequence of the resource in a group?

Regards,
Parkirat Singh Bagga.



--
View this message in context: http://linux-ha.996297.n3.nabble.com/Resource-Collocation-v-s-Resource-Groups-tp14699.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

Re: Heartbeat haresources with IPv6

Digimer — 2013-06-17T17:20:41

Ho Thiago,

   Heartbeat is deprecated and has not been developed in some time. 
There are no plans to restart development, either. It is _strongly_ 
advised that new setups use corosync + pacemaker. You can use the IPv6 
resource agents with it, too.

   The best place to look is on clusterlabs.org's "Cluster from Scratch" 
tutorial. It covers as the first example setting up an (IPv4) virtual IP 
address. It should be easy to adapt that to your IPv6 implementation. 
You will see two versions; One for crmsh and one for pcs. I would 
recommend the crmsh version for Ubuntu.

Cheers

On 06/17/2013 11:35 AM, listas< at >adminlinux.com.br wrote:

Heartbeat haresources with IPv6

2013-06-17T15:35:06

Hi,


 I'm using Ubuntu 12.04 + Heartbeat 3.0.5-3ubuntu2 to provide high availability for some IP addresses.
 I want to configure an IPv6 address on my haresources. I did this:
 
 File /etc/heartbeat/haresources:
 
 server.domain.com \
 &nbsp;&nbsp;&nbsp; 192.168.2.62/32/eth1 \
 &nbsp;&nbsp;&nbsp; 192.168.2.64/32/eth1 \
 &nbsp;&nbsp;&nbsp; 192.168.2.72/32/eth1 \
 &nbsp;&nbsp;&nbsp; IPv6addr::2001:db8:38a5:8::2006/48/eth1 \
 &nbsp;&nbsp;&nbsp; MailTo::adm< at >domain.com
 
 The IPv4 addresses work fine, but I'm not getting success with the IPv6 address.
 My logs shows this message:
 ResourceManager[22129]: info: Running /etc/ha.d/resource.d/IPv6addr 2001:db8:38a5:8 2006/48/eth1 start
 ResourceManager[22129]: CRIT: Giving up resources due to failure of IPv6addr::2001:db8:38a5:8::2006/48/eth1
 ResourceManager[22129]: info: Running /etc/ha.d/resource.d/IPv6addr 2001:db8:38a5:8 2006/48/eth1 stop
 ResourceManager[22129]: info: Retrying failed stop operation [IPv6addr::2001:db8:38a5:8::2006/48/eth1]
 
 Apparently there is a conflict between the characters '::' inside the IPv6 address and the separator '::' used in the haresources. But I would not like have to expand the IPv6 address. 
 
 Does anyone know a way to avoid this conflict?
 
 Thanks!
 --
 Thiago Henrique
 www.adminlinux.com.br







_______________________________________________
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: Antw: Re: Pacemaker - Resource dont get started on the standby node.

emmanuel segura — 2013-06-17T10:25:00

Hello Parkirat

Thank you very much


2013/6/17 Parkirat <parkiratbagga< at >gmail.com>

Re: Antw: Re: Pacemaker - Resource dont get started on the standby node.

Parkirat — 2013-06-17T09:30:53

Thanks Ulrich,

I have figured out the problem. 
The actual problem was in the configuration file for the resource httpd. It
was correct in the Master node but the configuration was missing in the
standby node, which was not allowing it to start.

Regards,
Parkirat Singh Bagga.



--
View this message in context: http://linux-ha.996297.n3.nabble.com/Pacemaker-Resource-dont-get-started-on-the-standby-node-tp14686p14695.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

Re: Antw: ocf HA_RSCTMP directory location

Ulrich Windl — 2013-06-17T06:16:12

<206418282.11638415.1371219712940.JavaMail.root< at >redhat.com>:
[...]
[...]

But you only have to do it right once (in a procedure/function):
If the PID file exists
then
   if the PID file is newer than the time of reboot
   then
      if there is a process with this pid
        if the process having that pid matches a given pattern
        then
          the process is alive
        else
          another process has this PID; remove stale PID file
        fi
      else
        remove stale pid file
      fi
   else
     remove stale pid file
   fi
fi

Everything else ist just wrong IMHO.

Regards,
Ulrich

_______________________________________________
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

Antw: Re: Pacemaker - Resource dont get started on the standby node.

Ulrich Windl — 2013-06-17T06:07:49

Hi!

The problem seems to be this:
Jun 15 20:16:27 prod-hb-nmn-002 pengine: [4401]: WARN: unpack_rsc_op:
Processing failed op apache_start_0 on prod-hb-nmn-002: unknown error (1)

Check why apache won't start.

Regards,
Ulrich

<1371327853299-14687.post< at >n3.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

Re: Pacemaker - Resource dont get started on the standbynode.

emmanuel segura — 2013-06-16T21:00:46

Hello Parkirat

can you share with us what was the problem? maybe this can help others
persons

Thanks


2013/6/16 Parkirat <parkiratbagga< at >gmail.com>

Re: Pacemaker - Resource dont get started on the standbynode.

Parkirat — 2013-06-16T20:33:29

I figured out the problem.

Thanks and Regards,
Parkirat Singh Bagga.



--
View this message in context: http://linux-ha.996297.n3.nabble.com/Pacemaker-Resource-dont-get-started-on-the-standby-node-tp14686p14691.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