Get logs from down'ed vm?

[dschwoerer]

I am trying to figure out a random failure (random segfault). I tried running the test suite a few times.

I had one run where I saw a few failures, and when I came back later, the failures had disappeared.
Looking at the backend.log, I saw that the VM went unresponsive and a new VM has been used.

Is there some way to get the last log from the first VM?
Maybe even some additional logs?

This is the backend.log and this the link to the build.

Logs would be appreciated :-)
If that can crash a VM, I am not eager to test this locally.

[FrostyX]

Is there some way to get the last log from the first VM?
Maybe even some additional logs?

Unfortunately no, the builder machines are removed.
But this may help you debug the issue https://frostyx.cz/posts/ssh-access-to-copr-builders

[nikromen]

Looking at the backend logs I can see that the VM booted successfully after a few attempts and the build was successful. What I can see is that the ssh to the VM failed to 54.208.45.94 which is - to my knowledge - VM in aws. After some retries, the aws VM was discarded and it got new one 2620:52:3:1:dead:beef:cafe:c193 (not aws). This really could be just aws issue - for these cases, we have the retry feature there. Can you confirm this was just one-timer or does it happen regularly?

[nikromen]

why do you think this is segfault?

closing as not a bug since the retry mechanism when ssh fails is a feature, but if you think I missunderstood you, feel free to reopen this