On Tue, Jul 24, 2012 at 04:01:40PM +0100, Aboubakr Seddik Ouahabi wrote:
> Hey there, I've created a thread somewhere, but I guess this is the right
> place to seek help for this, and here is my issue as stated there:
> 
> 
> Ok guys, that was very much appreciated and I thank you again. For now, I
> just want to get heartbeat to function as it should and I don't want to
> create a whole new thread for it.
> 
> As I said before, I have one public IP to access the server, and 2 nodes
> with 2 internal IPs, both are connected using Eth0, and what I want exactly
> is, if either one of Httpd or MySQL went down, the second node should take
> control and the virtual IP shall be assigned to it, until everything is in
> Sync again, then the primary or the favorible node should be taking over
> again.
> 
> Heartbeat is starting just fine, detecting the 2 nodes, then I tried to
> shutdown one of them and see what would it say
> 
>  Code:
> 
> cl_status nodestatus node02
> dead
> 
> And it found it was dead, but the failover isn't happening. I've tried to:
> 
>  Code:
> 
> service httpd stop
> 
> On node01, but it didn't switch anything to anything, so what I've been
> missing in my config? And here are the config I've tried in my ha.cf:
> 
>  Code:
> 
> # Logging
>  debug                          1
>  use_logd                       true
>  logfacility                    daemon
> 
>  # Misc Options
>  traditional_compression        off
>  compression                    bz2
>  coredumps                      true
> 
>  # Communications
>  udpport                        21xxx
>  bcast                          eth0
>  ucast          eth0            10.25.45.81
>  ucast          eth0            10.25.45.82
> 
>  autojoin                       any
> 
>  # Thresholds (in seconds)
>  keepalive                      1
>  warntime                       6
>  deadtime                       10
>  initdead                       15
> 
> crm respawn
> 
> node node01
> node node02
> 
> And I've tried 2 combinations for my cib.xml:

learn to use the crm shell, so much easier to the eyes...

> 
> 1:
>  Code:
> 
> <cib>
> <configuration>

I think you are missing no-quorum-policy=ignore


-- 
: Lars Ellenberg
: LINBIT | Your Way to High Availability
: DRBD/HA support and consulting http://www.linbit.com
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems

Reply via email to