1451835 - Deploy Windows 10 moonshot nodes in MDC2

Status: RESOLVED FIXED

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

Comment 1

[Mark Cornmesser [:markco]]

The plan is to do this post moonshot firmware upgrade. 

jmaher: Should we set up a staging pool for you all to take a look at before doing the complete deployment to MDC2?

Comment 2

[Joel Maher ( :jmaher ) (UTC -8)]

I don't think we need a staging pool as the firmware/osinstall will be identical.

Comment 3

[Bogdan Crisan [:bcrisan] (EEST - GMT + 3)]

Hello, as discussed, I just finished re-image the 8th chassis and I want to report a few machines that still don't want to appear in taskcluster and take jobs: 318, 320, 321, 322, 323, 324, 345.

The above mentioned machines have been re-image twice, after the first re-image I waited like 30 minutes to take jobs, (atm the pending tests are more then 900) meanwhile the others took more at least 3-4 jobs and completed them without exceptions

Comment 4

[Bogdan Crisan [:bcrisan] (EEST - GMT + 3)]

Continued with the re-image process, all of the machines from chassis 9 are continuing PXE booting, except for MS-361 which did not had that problem and 377 which has no display output.

So far I tried rebooting the machines and booting from the SSD/HDD, the install continues but at the next reboot the machines PXE boots again.

I also tried to modify the boot order, but it did not make any difference.

I'll try to see if the machines from chassis 10 are behaving the same.

Comment 5

[Bogdan Crisan [:bcrisan] (EEST - GMT + 3)]

And the same booting behavior can be seen on chassis 10.
At this point, I stopped re-imaging them since I cannot get them up and working, also shutdown the ones I worked on (all of the machines with PXE booting from chassis 9 and 10 (the first six))

Comment 6

[Mark Cornmesser [:markco]]

Because those have not been imaged previously, they will need to have the node boot order set to m.2, pxe from the chassis.

Comment 7

[Adrian Pop]

I've setup chassis 9 for windows machines to that the boot order : m2, pxe. Will check with the first 3 machines, if the issue still persists and report here

Comment 8

[Adrian Pop]

After changing the node boot order, I was able to reimage almost all of the machines from chassis 9

Comment 9

[Radu Iman[:riman]]

I have finished to re-image all the machines from chassis 10. No issues encountered. Waiting for them to take jobs. We'll go further with chassis 11.

Comment 10

[Radu Iman[:riman]]

Updates: 
   I've re-imaged the following machines once again: 318, 320, 321, 322, 323, 324, 345 . The process seems to be successfully finished but the workers are still missing from TC.
   The rest of the re-imaged machines from the chassis 8, 9 and 10 are working well.

Comment 11

[Roland Mutter Michael (:rmutter)]

Seems like chassis 11 wasn't a lucky one. Proceeded to reimage the first 12 machines and 11 of them have the pxe boot problem. Skipping 11 I've manage to finish reimaging chassis 12 . It's going to need a check later on , but for now, machines started picking up jobs.

Comment 12

[Adrian Pop]

I have reimaged chassis 11, 13 and 14. I had some issues with some of the machines. I have created bugs for these machines :
T-W1064-MS-599, T-W1064-MS-600, T-W1064-MS-474, T-W1064-MS-471