414400 - need a talos mac 10.4.8 image

Status: RESOLVED FIXED

(mozilla.org Graveyard :: Server Operations, task)

Description

[alice nodelman [:alice] [:anode]]

The image for talos mac currently on qm-ref-tiger is for 10.4.10.  The production talos mac tiger machines are on 10.4.8.  I didn't notice the difference when I was putting the image together.

The fastest way, as far as I can figure, to get to the correct state is to take an image of qm-pmac01 and put it on qm-ref-tiger.  Then I'll log on to qm-ref-tiger, clean up the state of the machine to remove references to qm-pmac01 and then we can re-image.  Then we should have a clean mac 10.4.8 image.

So, step one: grab image of qm-pmac01 and install it on qm-ref-tiger.

Comment 1

[Phong Tran [:phong]]

I'll grab an image of qm-pmac01 tonight.  I'll let you know when I get the image on qm-ref-tiger.

Comment 2

[alice nodelman [:alice] [:anode]]

Can you give me warning before you do so?  qm-pmac01 is currently running talos tests so I'll want to get it back up and running as quickly as possible after we've taken the image.

Comment 3

[Phong Tran [:phong]]

What's the best way to contact you?  I should be able to get to the data center by around 7 PM.

Comment 4

[alice nodelman [:alice] [:anode]]

I'm going to tell people in advance that the machine is going down, so as long as you email when me when it is back up and reachable that should be fast enough.

Comment 5

[Phong Tran [:phong]]

Let me know when I can shut this box down to take the image.

Comment 6

[Dave Miller [:justdave]]

Can I assume that the page I just got telling me that qm-pmac01 is down is related to this?

Comment 7

[alice nodelman [:alice] [:anode]]

Yes, this is the reason that qm-pmac01 is currently down.

When will the machine be back online?

Comment 8

[Phong Tran [:phong]]

qm-pmac01 is up and running again.

Comment 9

[Phong Tran [:phong]]

qm-ref-tiger imaged with image taken from qm-pmac01.

Comment 10

[alice nodelman [:alice] [:anode]]

I think that qm-ref-tiger is now ready for re-imaging.

Once we have that image can we apply it to qm-pmac05?  I can then use that machine to ensure that the image is correctly configured.

Then we can stop the whole machine-image-machine-image dance.

Comment 11

[Phong Tran [:phong]]

What's the username/password for qm-ref-tiger?

Comment 12

[Phong Tran [:phong]]

I am unable to logon with the username and password provided.  Can you check to make sure there isn't something wrong with this image?

Comment 13

[alice nodelman [:alice] [:anode]]

Now I can't reach the box at all.  Is it down or frozen?

Comment 14

[Phong Tran [:phong]]

The box should be online.  I am able to ping it from qm image.  If you can't get to it, then I will have to re-image it with the qm-pmac01 image and have you clean it up again.

Comment 15

[Phong Tran [:phong]]

I've put the qm-pmac01 image on qm-pmac-fast01 (10.2.73.220).  Please clean that up and I will recreate an 10.4.8 image based on that.

Comment 16

[alice nodelman [:alice] [:anode]]

I'd really rather not touch qm-pmac-fast01 anymore (since that'll slow me down in getting it up and on to stage).

Let's apply the qm-pmac01 image to qm-ref-tiger again.  I can have another go at it and see if I can get it configured without breaking anything.  This way I can still go ahead with my work on the fast machines while we are figuring out the last of the mac image issues.

Also - if you are going to be at the colo anyway, let's get the qm-pmac01 image onto qm-pmac05.  Then I won't be blocked on that machine either.

In conclusion:

- consider qm-pmac-fast01 done and complete
- apply qm-pmac01 image to qm-ref-tiger
- apply qm-pmac01 image to qm-pmac05

Comment 17

[Phong Tran [:phong]]

I'll apply the qm-pmac01 image to qm-ref-tiger and qm-pmac05.  Once you have qm-ref-tiger cleaned up.  I'll take an image of that and test it on one of the spare machine to make sure it works.

Comment 18

[alice nodelman [:alice] [:anode]]

Cool - this all sounds like a good plan.

Comment 19

[Phong Tran [:phong]]

qm-ref-tiger & qm-pmac05 re-imaged.  Let me know when qm-ref-tiger is ready to be used as a master 10.4.8 image.

Comment 20

[alice nodelman [:alice] [:anode]]

In my attempts to get qm-pmac05 and qm-pmac-fast01 reporting I've found that firefox will install but won't run on either of them.  This is with builds that can be installed/run on other machines.  I'm thinking that the image is damaged in some way, or the installation of the image has gone wrong.

I'm really at something of a loss as to what could be going wrong.  We can re-apply the image but I don't know if that will give us any additional information as to what isn't working.

The behavior that I see is that the browser doesn't open, but I see the firefox-bin process taking up 100% of my cpu.  I've looked at ktrace and gdb, but since I'm not working with debug builds I can only get a limited set of information.

Comment 21

[Phong Tran [:phong]]

Can we try to pull the image off qm-pmac01 again?

Comment 22

[alice nodelman [:alice] [:anode]]

We can try that, yes.

Comment 23

[alice nodelman [:alice] [:anode]]

Do we have a time line on the next imaging effort?

Comment 24

[alice nodelman [:alice] [:anode]]

I take it all back.  Both qm-pmac-fast01 and qm-pmac05 have decided to start working.  We can leave both of those machines alone.

After all of this, I'd like to just consider the current image of qm-pmac01 to be the master image.  We can put it on qm-ref-tiger and take a new image if anything changes - but for now it is good in its current state.

Comment 25

[Phong Tran [:phong]]

That image is already on qm-ref-tiger.  Please verify that this is working there as well.

Comment 26

[alice nodelman [:alice] [:anode]]

qm-ref-tiger is having display issues (appears all smudged and broken up into lines).  This probably can't be diagnosed without attaching a monitor to it.  So, the qm-pmac01 images appears to be working everywhere except on qm-ref-tiger.

Comment 27

[Phong Tran [:phong]]

I plugged a monitor into qm-ref-tiger and it came up fine for me.

Comment 28

[Justin Fitzhugh]

verified that the screen is at the right resolution and works via vnc.