TOPIC:

Hi I have created the test environment and when I stop master the full pool goes off line. That does not seem right?
My expectations would be that slave takes over.


Then when I bring the master up the pool comes up but the status of iscsi for slave shows as unplugged.
What am I doing wrong?

Hi,
Hope that someone will help ...
I have redone the installation and read everything I could find and I am having the same issue.
In the logs on the slave ( when master goes down ) I can see this:

Sep 5 13:57:07 xs2 ha-lizard: 14863 Failed to update global pool configuration settings in /etc/ha-lizard/ha-lizard.pool.conf - Check Configuration!


Sep 5 13:57:07 xs2 ha-lizard: 14863 Mail Spool Directory Found /dev/shm/ha-lizard-mail
Sep 5 13:57:07 xs2 ha-lizard: 14863 email: Duplicate message - not sending. Content = Failed to update global pool configuration settings in /etc/ha-lizard/ha-lizard.pool.conf - Check Configuration!
Sep 5 13:57:07 xs2 ha-lizard: 14863 email: Message barred for 60 minutes
Sep 5 13:57:07 xs2 ha-lizard: 14863 This host has self fenced - HA forcefully disabled. Exiting...
Sep 5 13:57:07 xs2 ha-lizard: 14863 HA can be reactiveted for this host by running /etc/ha-lizard/scripts/recover_fenced_host

running recover script does nothing .....
ha-cfg status hangs on slave

after I bring the master up everything seems to go back to normal ...

Based on what I read ( and I may be wrong ) the slave should take over.
Any help would be appreciated.

Your configuration seems fine.. Make sure that your slave can ping 192.168.199.1 otherwise it will fail to fence and takeover (become master).

Before running your test again, check status with "ha-cfg status" and make sure you are not in suspended mode.

Thank you for input.
I have verified it and ping works.



After master goes down slave reboots and this is what I get on the console of the slave. Please see attached.




I am running these two xenservers in virtualbox. Would that have anything to do with it?

before the slave goes down following are last lines of the ha-cfg log:

More lines from the log file are needed to know exactly what is going on. Can you perform your test again and post all the log lines from the slave from the moment the master goes away until the slave fails to fence the master.