1472541 - (T-W1064-MS-027) [MDC1] T-W1064-MS-027 problem tracking

Status: RESOLVED WONTFIX

(Infrastructure & Operations Graveyard :: CIDuty, task)

Description

[Adrian Pop]

Attachment: T-W1064-MS-027.png

The machine can't be found on Task Cluster.
We have tried rebooting it but this didn't solved the issue.
We have tried to reimage it but it gives an error, that can't be bypassed.
Please check attachment

Comment 1

[Attila Craciun [:arny]]

I have managed to start the re-image process and will update the bug once the server is back to business.

Comment 2

[Attila Craciun [:arny]]

T-W1064-MS-027 is back in TC, waiting for jobs.

Comment 3

[Attila Craciun [:arny]]

The server completing jobs fine now.

This issue appear because the partitions of the HDD are in a non FAT/NTFS format. When I connected to ILO at the fist time, I get the Windows login screen with user administrator asking for password. I have rebooted and I get Ubuntu booting! :|. Then I have tried several live cds (debian, slackware, slax, centos) to boot it and try to erase the partitions. No one of the lives succeeded to boot so I decided to boot via PXE several times. After a few try, I get another error where I had option to Retry or Cancel so I hit cancel and the installer opened me in a cmd, where, I used diskpart to erase the partitions.
Then, rebooted the server, selected network boot and the re-image was successful.

Is there a way to add a disk management option/image right at the PXE menu? This will help much if we get same errors. Or, update the installer to wipe any partitions before a new Windows install.

Comment 4

[:dhouse]

Mark, is there a way to make the windows netboot installer reformat the disk?

Comment 5

[Mark Cornmesser [:markco]]

As is it should be formatting on each new install. If the Windows is asking for the administrator password 2 things had happened. Either the deployment failed or the first run of OCC is still going. The best bet is to log in, we are working on getting the password into a repo. Once logged in, check the task manager and see if powershell if running. If it is then OCC may still be doing configuration work. If it is not running then it is safe to assume OCC had failed on the initial setup. I would then try an additional install. 

Regarding the HDD we are currently not using that drive on the nodes. M2 is the one we are concerned about.

Comment 6

[Attila Craciun [:arny]]

Attachment: ms-disks.PNG

I have re-imaged again since yesterday I put the old image, now is running the new generic one. 

Also, I made a capture with the disks, so, the 64GB (M2?) one is used for OS and the 250GB(SATA?) for what? When I erased the partitions yesterday, it was the 250GB one.

Comment 7

[Attila Craciun [:arny]]

Attachment: errorpng-dep-share.png

This was the error before droped me to cmd.

Comment 8

[Attila Craciun [:arny]]

I have re-images this twice today with the win 10 generic 10 and still not getting new jobs. Yesterday was fine with the win 10 image. Should we use the old image or do you need the win10 generic10 for tests?

Comment 9

[Zsolt Fay [:zfay]]

Also @markco should we reboot / re-image other win10 machines from chassis 1 if we see them misbehaving? I'm guessing you can't test the new generic worker if they don't run tests.

Comment 10

[Mark Cornmesser [:markco]]

(In reply to Zsolt Fay [:zsoltfay] from comment #9)
> Also @markco should we reboot / re-image other win10 machines from chassis 1
> if we see them misbehaving? I'm guessing you can't test the new generic
> worker if they don't run tests.
Reboot and reimages are fine.

Comment 11

[Radu Iman[:riman]]

This worker is running jobs. I'll close the bug for now.
https://tools.taskcluster.net/provisioners/releng-hardware/worker-types/gecko-t-win10-64-hw/workers/mdc1/T-W1064-MS-027

Comment 12

[Roland Mutter Michael (:rmutter)]

Looks good now : https://tools.taskcluster.net/groups/K1aHi1obRGmPlSqWQsTS_A/tasks/Hce4ulDfR36ERDnjjifAIA/runs/0 . Closing for now.

Comment 13

[Adrian Pop]

re-opening the ticket. The worker hasn't received any job for 5 days.
I've tried to reimage it but I wasn't able to boot it from the network.
XGE1/0/27 DOWN auto A A 240
The interface is down and it will be needed to be reseated.

Comment 14

[Adrian Pop]

following the last comment https://bugzilla.mozilla.org/show_bug.cgi?id=1369661#c4 from bug 1369661, this machine has been decommissioned. I will close the bug.