676518 - reboot requests (scl1)

Status: RESOLVED FIXED

(Infrastructure & Operations :: RelOps: General, task)

Description

[[e-mail] (:bhearsum)]

talos-r3-w7-001

Comment 1

[Zandr Milewski [:zandr]]

talos-r3-w7-001: powered off. Got the "Windows Error Recovery" screen on boot, and is now running the "Piriform Defraggler"

Comments welcome.

Comment 2

[William Lachance (:wlach)]

Yes, that's expected. I was testing whether defragmenting would have any impact on the very bad disk i/o results we were getting on this machine.

Could you see where the machine is at? talos-r3-w7-001 still does not seem to be reachable on the build-vpn network.

Comment 3

[Zandr Milewski [:zandr]]

Probably early next week before I'll be back in that datacenter.

Comment 4

[Zandr Milewski [:zandr]]

And... for that matter, a simple reimage would have achieved the same thing.

Not sure what the goal here was, nor why it wasn't simply filed with ops as a separate bug. Fragmentation isn't going to be the problem. A dying hard disk is.

Comment 5

[William Lachance (:wlach)]

Zandr: Have a look at bug 675363. The general goal is to figure out why some xpcshell tests seem to be performing so poorly on these machines, so we can make them go faster.

It's possible there's an issue particular to that machine's hard disk, yes. It's not the only possibility though.

At the time, we were just trying to test out various theories about why things were running so slowly. Unfortunately one of these experiments seems to have unintentionally brought the machine down.

Comment 6

[William Lachance (:wlach)]

This should be reopened. Note that this machine being down isn't blocking work, so no need to rush on this. Next week should be fine.

Comment 7

[Dustin J. Mitchell [:dustin] (he/him)]

reboot bugs don't get re-opened.  Opened bug 676671.