patch to disable linux32 talos builders. I'll clean up the configs to remove that machine class on a subsequent patch/reconfig.

builder diff

From email conversation Armen suggests a pool allocation for the reimaged machines ====== I would say 70% win7 and 30% win8. Current backlog (which looks like this most of the time): win7-ix (2029) win8-ix (1198) winxp-ix (86) ======

Comment on attachment 8666118 [details] [diff] [review] bug1208449.patch Review of attachment 8666118 [details] [diff] [review]: ----------------------------------------------------------------- ::: mozilla-tests/config.py @@ -207,2 @@ > NO_WINXP = [platform for platform in ALL_TALOS_PLATFORMS if platform != 'xp-ix'] > NO_OSX = get_talos_slave_platforms(PLATFORMS, platforms=('linux', 'linux64', 'win32', 'win64')) we have another instance of 'linux' here.

builder diff

Comment on attachment 8666680 [details] [diff] [review] bug1208449.patch Review of attachment 8666680 [details] [diff] [review]: ----------------------------------------------------------------- Leaving the r? on joel, since I'd like a "written down" comment from someone on ateam in support of this plan. But it doesn't have to be joel, and doesn't have to be on this bug/patch. (so may already exist) The code itself however, looks great.

r+ from me as well! We are all sorting this out together!

in production since last reconfig at 8am PT. Will disable machines in slavealloc next

the 44 machines that were enabled are now disabled in slavealloc

add new range of windows test machines and decomm ubuntu32_hw class of machines

add new range of windows test machines and decomm ubuntu32_hw class of machines

Comment on attachment 8667497 [details] [diff] [review] bug1208449bb.patch Review of attachment 8667497 [details] [diff] [review]: ----------------------------------------------------------------- r+ because config is sane, but there is a gap here I think. ::: mozilla-tests/production_config.py @@ -16,5 @@ > > for i in range(1, 173): > SLAVES['xp-ix']['t-xp32-ix-%03i' % i] = {} > > -for i in range(1, 173): count(new_machines) == count(old_machines) + 31 @@ -21,3 @@ > SLAVES['win7-ix']['t-w732-ix-%03i' % i] = {} > > -for i in range(1, 102) + range(103, 181): # Omit 102 for win10 // Bug 1191481 count(new_machines) == count(old_machines) + 14 @@ -58,5 @@ > for i in range(1, 2100): > SLAVES['ubuntu64_vm']['tst-linux64-spot-%03i' % i] = {} > > -for i in range(1, 56): > - SLAVES['ubuntu32_hw']['talos-linux32-ix-%03i' % i] = {} count(removed_machines) > count(added_machines) What happens to the extra 11 machines?

machines to add vi dbimport

Comment on attachment 8667518 [details] bug1208449db.txt added to slavealloc, disabled

re comment #16. there aren't an extra 11 machines. I suspect the range wasn't cleaned up when we last moved things around. According to slave health there were only 45 machines in production until yesterday.

This can be closed, there are only a few machines left to be brought up in bug 1209492 (Q is working through reimaging issues)