TOPIC:

Hello:

We're currently running HA-Lizard 1.7.8 on XenServer 6.2 (two-node clusters). I was wondering if there are any guidelines or a checklist on how to re-install a failed host (either the master or the slave host), if it completely goes up in smoke. I've read through the documentation and I see mention of how it fails over, but I'm looking more for step-by-step on how to re-install the host and get it back up and running if it completely fails. I suppose I could follow the installation instructions but I was looking for something specific when there's a host failure.
Thanks for any recommendations!

--
Chris

Hi Chris,
We do not have a specific document for re-introducing a host. The procedure would be nearly identical to the steps in the reference design with one major exception.

Since the storage already has real data - you would perform the DRBD setup on the new host ONLY. Meaning - do not format, initialize or create metadata for the current storage. Also - when connecting the DRBD secondary device pay careful attention to ensure that the new device is overwritten with the data (not the other way around). Much of this is well documented in the DRBD community. If you happen to perform such a re-installation we would be glad to publish a document if you supply your tested procedure.

OK thank you for the followup. Your suggestions on setting up DRBD on the new host makes sense, and that is my main area of concern. Unfortunately I do not have any hardware to run through a test scenario as of right now, which was my main reason for inquiring as well as getting an official answer. Thank you for your help. If I am able to run through the failure scenario I will definitely post this back.

Best,
Chris