Security

Hello everyone,

In my company with have a cluster of 5 Clearpass servers with some latency between servers.

Yesterday we tried to upgrade the cluster manually from 6.7.13 to 6.8.5, I mean one server at a time. Not with the "cluster upgrade" function because of the latency between servers as I said.

However, when we did manual upgrade on the publisher from 6.7 to 6.8.5, the server's declustered and we lost all licenses so we rolled back.

Moreover, when we installed the firmware and rebooted the publisher, the server booted on 6.7.13 instead of 6.8.5. In order to boot on 6.8.5, we had to do it manually in CLI with the command : system boot-image.

A solution for this issue ? 

What do you recommand about this upgrade from 6.7.13 to 6.8.5 ? The best practises for the smooth running of the operation.

Manual upgrade on each node, or "cluster upgrade" ?

Thank you in advance for your time.

Regards,

Rayane

Hi Rayane,

I believe, the first step of your upgrade is from 6.7.13 to 6.8.0. This upgrade should set the boot image to 6.8.0 partition upon the installation and ready to restart the server. If this is not happening, then there is something went wrong with the upgrade and need to be analysed.

Please open a TAC case with the server logs from 6.7.13 to go through the upgrade logs and derive the action plan for next upgrade.

Hello Saravanan,

Thank you for your quick response.

Yes, that's exactly what we did via the installer on Clearpass.
Strange thing is we went straight to version 6.8.5 and not 6.8.0. (see pic)

I will open a case for log analysis.

Thank you,

Rayane