TOPIC:

A few more things to consider:

- Stopping the halizard service ("service ha-lizard stop") or preventing it from starting at boot ("chkconfig ha-lizard off") will not fully stop it unless the watchdog is also stopped and prevented from starting at boot ("service ha-lizard-watchdog stop" and "chkconfig ha-lizard-watchdog off")

- Any chance that you have also enabled XenServer HA from within XenCenter? If so, HA-Lizard will detect this and disable itself. The is really the only way that I see XenServer interfering in your test case. IF XenServer HA is definitely off, then the issue is likely within HA-Lizard.

A dump of /var/log/messages from the time the server boots would be very helpful.

Just for clarification and my better understanding:

The pool HA Status is disabled.

The pool OP_MODE is set to 1

The four VM do have HA-Enabled true

all VM are not running. If I start manually(!) in XenCenter the vApps group the should come up in the right order and delay.

If I remove them form their homeserver and repeat the manual start of the vApps group they start all at the same time.

Does this behaviour really interfer with HA-Lizard?

If I disable the Watchdog service too, should this change something. I could also send you the logs if you tell me whom to address.

BR Andreas

Thanks for the clarification. I now better understand the issue and agree, this is not ha-lizard related. FYI - we tested with halizard completely turned off and see the same behavior you are describing - the delay is not being respected between the VM starts.