Comments on: Howto setup a haproxy as fault tolerant / high available load balancer for multiple caching web proxies on RHEL/Centos/SL

By: robert

robert — Thu, 08 Nov 2012 17:58:36 +0000

Hi!

The version is quite critical … thats the reason I talked about rhel6 … they also may changes in the service packs e.g. the non-multicast feature is quite new.

By: phillip

phillip — Thu, 08 Nov 2012 17:56:35 +0000

addition to 15 and 16.
http://www.gossamer-threads.com/lists/linuxha/pacemaker/71761

make sure you upgrade Pacemake <= 1.1.2 so that it will work

By: phillip

phillip — Thu, 08 Nov 2012 15:05:04 +0000

addition to the question 15… it would seem that the pacemaker did not install correctly.

By: phillip

phillip — Wed, 07 Nov 2012 19:59:49 +0000

I am following your document and I ran across an issue. when I try to start the pacemaker service is it not working.I am good up to that point. I also tried “service pacemaker start” .. no good. any tips??
I am installing pm 1.0.12-1.e15 on RHEL5

By: Lior

Lior — Wed, 31 Oct 2012 12:41:49 +0000

when i do this part:
crm(live)# configure show

i only get:
node localhost

instead of:
node proxylb01.int
node proxylb02.int

why is that?

By: robert

robert — Sat, 15 Sep 2012 11:16:07 +0000

1. I see only one node configured … configure the second too: “1 Nodes configured, 2 expected votes” .. you need to do the first steps on both nodes.

2. take a look into the
/usr/lib/ocf/resource.d/heartbeat/haproxy file and search for the monitor section, try the steps by hand if they work or a program is need and not installed.

By: Steve

Steve — Wed, 12 Sep 2012 21:39:02 +0000

Have followed this guide and have a couple of issues.

First when I issue the crm_mon -1 command I only see one node listed not two. The same is true for issuing the configure show command. Also I am getting an error with the crm_mon -1 command.

============
Last updated: Wed Sep 12 16:27:02 2012
Last change: Wed Sep 12 14:40:16 2012 via cibadmin on proxy01
Stack: openais
Current DC: proxy01 – partition WITHOUT quorum
Version: 1.1.7-6.el6-148fccfd5985c5590cc601123c6c16e966b85d14
1 Nodes configured, 2 expected votes
2 Resources configured.
============

Online: [ proxy01 ]

Resource Group: haproxyIPs
haproxyIP03 (ocf::heartbeat:IPaddr2): Started proxy01
haproxyLB (ocf::heartbeat:haproxy): Started proxy01

Failed actions:
haproxyIP03_monitor_0 (node=proxy01, call=2, rc=-2, status=Timed Out): unknown exec error

By: robert

robert — Tue, 07 Aug 2012 07:45:45 +0000

If you’re using the crm CLI it does error checking. you can also edit it and than commit it (press tab tab in the configure mode) … just dont use crm directly from the bash with the commands as parameters as than there is no error checking …

By: A.Genchev

A.Genchev — Tue, 07 Aug 2012 07:41:56 +0000

is it possible to work on a config file instead of “crm configure”, because I find the syntax error prone. For example “…params conffile=/etc/haproxy/haproxy.cfg op monitor interval=30s”
How the cluster manager will decide which are params for the ‘primitive’ and which are parameters for cman – probably by the magic word “op” it starts parsing own parameters ?

found an article(scientific lol) on this topic: http://ijaest.iserp.org/archieves/12-Jn-1-15-11/Vol-No.7-Issue-No.2/7.IJAEST-Vol-No-7-Issue-No-2-Autonomic-Fault-Tolerance-using-HAProxy-in-Cloud-Environment-222-227.pdf

By: Jared

Jared — Wed, 01 Aug 2012 10:53:17 +0000

The solution to:

crm(live)configure# primitive haproxyLB ocf:heartbeat:haproxy params conffile=/etc/haproxy/haproxy.cfg op monitor interval=30s
lrmadmin[28965]: 2012/08/01_10:51:06 ERROR: lrm_get_rsc_type_metadata(578): got a return code HA_FAIL from a reply message of rmetadata with function get_ret_from_msg.
ERROR: ocf:heartbeat:haproxy: could not parse meta-data:
ERROR: ocf:heartbeat:haproxy: no such resource agent

Was correctly setting the file permission (to 755) for the haproxy script.

-rwxr-xr-x 1 root root 5297 Aug 1 10:26 haproxy