Thanks! Will look at it around 2am UTC (7pm PDT)

I think that the current Linux Perf slaves stopped processing requests from chromium.perf master ~20 minutes ago:
https://build.chromium.org/p/chromium.perf/builders/Linux%20Builder

We might want to consider to restart chromium.perf master now, or submit a partial revert of https://codereview.chromium.org/1999793004/diff/1/masters/master.chromium.perf/slaves.cfg (the part that removes 5 existing slaves). In the latter case, a restart of those old slaves might be required.

Not sure why (may be some change outside of my view), but it's all good now. The slaves started to accept the requests.

*shrug* The slaves never restarted: http://shortn/_EV5rA2FPvf
so they didn't pick up the change yet, so I'd expect them to work. I'm not sure why they flickered.

BTW, what happens to the original slaves: 'slave21-c1', 'slave22-c1', 'slave48-c1', 'slave49-c1', 'slave52-c1'?

Are you going to use them anywhere, or should we recycle them?

we should recycle them. I keep list of the slaves went out of business. There's also vm853-m1, which probably needs to be destroyed.

Anyway, I will file a separate bug for that.

Re: #c5: cool, thanks!

Apparently, only slave{79,89,103}-c1 didn't reconnect properly. Restarting those.

All new slaves are online.

Thanks!