Re: GitLab partial outage - attempting to mitigate

Ben Gamari writes:

Bryan Richter via ghc-devs writes:

...

I eventually resorted to a server reboot, which cleared up all the problems I was seeing. I think we're back in business.

The root partition was close to running out of disk space yesterday. The problem appears to be that /nix is located on the small system drive. We should really address this although moving /nix is sadly not easy and will certainly require downtime.

In the meantime, I have significantly reduced the number of snapshots retained in the root dataset. This brought disk usage down from 70% to 12%, which should keep us afloat for a long while. Cheers, - Ben