New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
FreeBSD 12.0 jails dying #89
Comments
The problem of repeating jail IDs could be addressed by using random jail IDs in collisions, however; my use case requires repeatedly using the same experiment ID. The underlying problem is FreeBSD 12 (stable) doesn't complete deleting jails upon experiment termination. |
I've been digging a bit on my FreeBSD 12.0 virtual machine so I just wanted to write an update on this issue... First of all, this is definitely primarily a bug with FreeBSD jails (even without IMUNES). I can replicate it with:
There shouldn't be any, but there are dying jails left when I run I accidentally found out that the dying jails disappear when copying/creating a big file on my VM, for example:
This also works with IMUNES dying jails. Of course, this isn't a fix or anything, I just wanted to acknowledge this 'hack' (because this and this tells us that there is no fix for this, only that 'keeping the same JID or ... recycling freed JID is not recommended' and to just 'ignore it'. :) This and again this suggest that there is a problem with credentials holding a jail.
This seems like a correct solution to the IMUNES issue. I will look into this but it could be a bigger challenge for me to test and make sure I don't break anything so it could take a while. |
This patch resolves the issue, and jails now don't remain in the dying state: WARNING: I am not a FreeBSD developer and this is not an official patch, but according to this it should be ok to revert this one line to an older revision. How to patch:
|
In FreeBSD 12.0, sometimes jails cannot be destroyed and removed from the existing jail list, but rather stay in the "dying" state (this can be seen with
jls -ad
). This shouldn't be a problem except when trying to restart the experiment, when jails are destroyed and created with the same ID.We could bypass this issue by matching the current experiment ID with the running jail IDs before running every experiment and changing it if it already exists.
The text was updated successfully, but these errors were encountered: