I might have overlooked the information but maybe someone else has the same questions, so here it goes:
Is there a compatibility Matrix of how Checkmk versions 1.6 and 2.0 can interact?
I think I have already read that the 1.6 agent talks to a 2.0 server although features might be missing.
But what about server to server? Say I want to update my distributed setup, do I need to shut everything down and update simultaneously or can I do this in a rolling manner or can I even run a satellite for several weeks on an older/newer version?
Maybe someone can give a brief explanation on this. Thanks!
i assume (hope ) that the documentation is going to updated with some upgrade guides in the next days when checkmk is released. Additionally with a very high probability there will in depth talks at the next conference.
But reading the following werk i expect migrations from 1.6 to 2.0 to be easier than before, allowing different upgrade scenarios.
Yeah, upgrade guides would be quite nice, although in my experience Checkmk was always painless and dead simple to upgrade, so my main concern is compatibility within the upgrade process and rather shortly after.
I am waiting for quite some time now for some features in 2.0 so I want to upgrade really soon after the release. That’s why waiting for the conference is not my preferred scenario.
But maybe someone with deeper insight can add a little confidence here.
I will try to answer this with the experience from major upgrades before.
A distributed master-slave setup must be upgraded as one big upgrade.
What i have done before is the following and for 2.0 it would be the same.
First step forbid any changes on the master WATO. Upgrade all slaves one after another.
Livestatus should also work without problem between the versions. So no need to shutdown the complete system. As the last system i upgrade the master and allow changes again.
Another way is to upgrade the master first and do all slaves then. Here you need to pay attention that changes are only activated on the already upgraded slaves. If you are the only main admin on the system you can also do this way. What can happen here (this was before 1.4 to 1.5 i think) is that upgraded main site complains a little bit about missing livestatus features in the old slaves. Before there where changes in livestatus tables.
Alright, putting all the answers together I think my initial question can be answered as follows:
Checkmk will be compatible between version 1.6 and 2.0. However it is strongly recommended to not stay in that state. It should merely exist during the upgrade process.
I will update this after I performed the upgrade in my environment. Anyone able to share their experiences feel free! Maybe someone else is looking for this and can gain confidence from this post.
I cannot edit my anser anymore so posting this right below:
I upgraded our distributed environment the other day from CME 1.6.0p22 to CME 2.0.0p3 without any problems. So I can confirm, that during the upgrade process 2.0 and 1.6 were able to communicate through livestatus and after all hosts where on 2.0 the configuration sync worked like a charm.
So to anyone reading this: Go for the update, it is totally worth it! Just follow the docs linked above thoroughly.
This topic was automatically closed 365 days after the last reply. New replies are no longer allowed. Contact @fayepal if you think this should be re-opened.
) that the documentation is going to updated with some upgrade guides in the next days when checkmk is released. Additionally with a very high probability there will in depth talks at the next conference.