I was wondering what people's opinions are on running the CR repository on production systems? I'm currently predisposed to not do it since it indicates it hasn't fully gone through QA yet. But from my understanding except for the alt-arch systems these are just recompiles of the redhat RPMs after debranding right?
I'm Curious how other people are treating this repository since, from the sounds of it, rpms are going to be in holding here between releases.
So would you treat it as a production repository or for testing only? Do we know how it's actually intended to be treated?
Any opinion is appreciated.
On 9/12/19 3:10 PM, Gwaland wrote:
... So would you treat it as a production repository or for testing only? Do we know how it's actually intended to be treated?
Well, like virtually everything else, It Depends (TM).
The first depends, for me, is how critical are the fixes relative to what I need in production?
So, what I do, is keep a test server running on the same hardware as my production servers, and update to CR on it first, making sure it reboots ok, since I have had the reboot not be ok on a new 7.x release (the UEFI vars memory overflow due to shim rebased to v15; see https://bugs.centos.org/view.php?id=15522 ). Once I see it reboots ok, I test. Since my own laptop is running CentOS 7 as well, I'll update it to CR for testing; nothing like living it for your daily main computer to tune you in to issues!
Once I'm satisfied CR is stable, then, if and only if the first depends is met, I'll update affected servers to CR, if the fixed issues are critical enough. If they're not that critical, then I hold off until GA.
The second depends is whether there are other dependent packages I need that aren't yet updated. If so, then I wait until those packages are ready, then update.