Hey all,
So I’ve done a lot of yum updates with ansible and several times I get issues where partial updates are run and it says it’s been completed, however, I reboot and I’m in a kernel panic because the new kernel didn’t update properly so I have to revert to the previous kernel and remove the newest and reinstall it again. I also often wind up with transactions that haven’t been completed. Has anyone experienced this before?
Thanks!
Chris
what version of ansible are you using?
Using the latest of 2.7.
I've seen this also. When the new kernel installs there is a post
transaction scriplet to build the initrd. In some cases it doesn't
build correctly (I see it fail maybe 1-5% of the time) and you'll get
a panic when the system reboots. I don't know that I've seen enough
to blame it on Ansible, though.
--Steve
I had a similar issue before with centos but, it was more like a limit ( for example 4 ) of kernels that could be install at the moment but, the current running kernel was uninstalled and the new one did not install correctly so the system hanged during booting.
So does anyone have any recommendations on how to alleviate this problem? Perhaps some sort of delay before exiting or something?
well like mentioned this is not ansible specific so i suggest you research more into the OS side.
Well if I run it manually, I don’t have any issues. I’ll look into it further. Thanks again.
I know this is an ansible group but you could attach dmesg output if you get one after the kernel fails to boot. Check /var/log/boot-(date) /var/log/dmesg /var/log/yum.log …you can start from there.
Also I don’t know if you’re logging ansible actions every time, but you could find answers there.