667024 - Re-image some rev3 minis as win7 64-bit test slaves

Status: RESOLVED INVALID

(Release Engineering :: General, defect, P3)

Description

[Chris Cooper [:coop] (he/him)]

This should be done after the win64 build machines are online (bug 647287), and the 5 existing win64 test minis look green enough to be worth it.

Comment 1

[Matthew Larrain[:MaRu]]

We have 5 machines in production and 4 more about to go into staging for testing and waiting on bug 668521. I would like to get a single rev3 mini to start testing on it to verify how well it works also can these PXE boot properly.

Comment 2

[Zandr Milewski [:zandr]]

Disregard digipengi's last, this bug is talking about Talos machines, not builders. 

We'll stick with our existing win7 refimage for r3 minis, and worry about WDS for Win7 testers on whatever becomes the rev4 win7 Talos machine. (which will not be a mini)

Comment 3

[Armen [:armenzg]]

FTR we won't be re-imaging a full size pool but perhaps 5-10 slaves.

We would need to take another snapshot of t-r3-w764-ref since I deployed drivers and directx changes in bug 671260.

Comment 4

[Armen [:armenzg]]

(In reply to comment #3)
> We would need to take another snapshot of t-r3-w764-ref since I deployed
> drivers and directx changes in bug 671260.

I had forgotten to file this. Now on bug 667024.

Comment 5

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

Correction: bug 674607.

Comment 6

[Armen [:armenzg]]

Bringing back to Releng until our pool's capacity/state allow us to reimage win7x64 test slaves.

Nevertheless bug 674607 still stands to be done.

I will take this bug until we are ready.

Comment 7

[Chris Cooper [:coop] (he/him)]

Assigning to Armen based on comment #6.

Comment 8

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

This is still blocked on Armen.

Comment 9

[Chris Cooper [:coop] (he/him)]

(In reply to Dustin J. Mitchell [:dustin] from comment #8)
> This is still blocked on Armen.

To be clear, Armen is actually waiting on the deployment of rev4 minis (bug 671415) as 10.6 testers (and subsequent perf rebasing) so we can reclaim some rev3 minis for Win7 64bit testing.

Comment 10

[Armen [:armenzg]]

I am trying to figure out on what are the plans for win64 as adding machines and jobs without being tier1 is dumb (as devs ignore the issues and hide the jobs).
We should know in the next 2-3 weeks at the latest (I hope).

Right now we would have to hold up half a load since there are only opt builds (there a bugs for the debug jobs to be green) and we might be able to get by with a pool of 20-30 slaves for m-c, m-i & try (or even less). We'll see as we know more what the plans are.

Comment 11

[John Ford [:jhford] CET/CEST Berlin Time]

(In reply to Armen Zambrano G. [:armenzg] - Release Engineer from comment #10)
> I am trying to figure out on what are the plans for win64 as adding machines
> and jobs without being tier1 is dumb (as devs ignore the issues and hide the
> jobs).
> We should know in the next 2-3 weeks at the latest (I hope).

Armen, do you know what the call was?  I remember seeing something to the effect that we are not going to make it tier 1 for a while.  If we aren't making w7 64 a tier 1 platform, should we mark this bug as a dupe of bug 683734 and distribute the r3 machines to linux, linux64, xp and w7?

> Right now we would have to hold up half a load since there are only opt
> builds (there a bugs for the debug jobs to be green) and we might be able to
> get by with a pool of 20-30 slaves for m-c, m-i & try (or even less). We'll
> see as we know more what the plans are.

Comment 12

[Armen [:armenzg]]

We're not going to re-purpose any rev3 machines as w764 slaves until product decides to steer such platform as a tier1 platform.
This won't happen for few months if it happens that early.