Don't remove mysql in CI

[ChrisDenton]

I think this works around the issue?

r? @Kobzol

[workingjubilee]

can't we just stop trying to remove MySQL specifically? I know we want images with less stuff on them but it doesn't seem like it'd save us that much space.

[ChrisDenton]

Maybe? But my motivation is to get CI moving again so I'd prefer not to risk it here. I leave that for someone else to investigate at their leisure.

[workingjubilee]

I certainly won't insist on my idea if this works out.

[Kobzol]

We tried using disable before, and it was also causing similar issues. I think that the safer bet is just to avoid uninstalling MySQL.

[ChrisDenton]

Ok, I can do that. Two votes to one, ha.

[workingjubilee]

kinda sucks that GH requires us to use images with so much bloatware but w/e

[ChrisDenton]

Done.

[Kobzol]

Will wait a minute if PR CI passes the free disk stage, and then I'll approve it.

[Kobzol]

Looks like it went through. Hopefully this will fix the issue.. Thank you!

@bors r+ p=101

[bors]

📌 Commit cb9cb1d has been approved by Kobzol

It is now in the queue for this repository.

[bors]

🌲 The tree is currently closed for pull requests below priority 100. This pull request will be tested once the tree is reopened.

[bors]

⌛ Testing commit cb9cb1d with merge fc147b4...

[bors]

☀️ Test successful - checks-actions
Approved by: Kobzol
Pushing fc147b4 to master...

[rust-timer]

Finished benchmarking commit (fc147b4): comparison URL.

Overall result: no relevant changes - no action needed

@rustbot label: -perf-regression

Instruction count

This benchmark run did not return any relevant results for this metric.

Max RSS (memory usage)

This benchmark run did not return any relevant results for this metric.

Cycles

This benchmark run did not return any relevant results for this metric.

Binary size

This benchmark run did not return any relevant results for this metric.

Bootstrap: 790.257s -> 791.712s (0.18%)
Artifact size: 347.74 MiB -> 347.74 MiB (0.00%)

[Kobzol]

After this PR, our ARM jobs suddenly got an hour longer:

I don't think that the PR itself is the cause, rather it seems like GitHub changed something (which explains the sudden breakage of our CI), which might have also caused the ARM runners to become slower. I wonder if it was the rollback of the faulty ARM hardware (#135867). In any case, something to watch out for.

CC @marcoieni, as you might be interested in this.