685242 - split try and non-try tegra pools

Status: RESOLVED FIXED

(Release Engineering :: General, defect)

Description

[Aki Sasaki (not active)]

Tegras are multi-colored right now.
One hypothesis is that try runs are leaving the tegras in a fragile state.
If we can split the tegras up into try- and non-try- it may make us more stable.

https://tbpl.mozilla.org/?tree=Mozilla-Inbound&usebuildbot=1&jobname=Android
https://tbpl.mozilla.org/?tree=Firefox&usebuildbot=1&jobname=Android
https://tbpl.mozilla.org/?tree=Try&usebuildbot=1&jobname=Android

Comment 1

[Aki Sasaki (not active)]

Attachment: non-functional attempt #1, configs

Comment 2

[Aki Sasaki (not active)]

Attachment: non-functional attempt #1, custom

Spent several hours trying to change one thing and failing.
Still wishing for non-programmatic configs.

Comment 3

[Aki Sasaki (not active)]

Actually, this patch works.
dump_masters.sh doesn't output slave names.
Cleaning up the patches and asking for r?

Comment 4

[Aki Sasaki (not active)]

Attachment: split try and non-try tegra pools, buildbotcustom

Let me know if this is the right approach.
When I alter dump_master.py to print b['slavenames'], then I see the separate tegra pools.

Comment 5

[Aki Sasaki (not active)]

Attachment: split try and non-try tegra pools, buildbot-configs

configs portion.

Bear, let me know if the split makes sense. Ideally this is the first stab at it and we can adjust the pools at will.

Comment 6

[Mike Taylor [:bear]]

Comment on attachment 559020 [details] [diff] [review]
split try and non-try tegra pools, buildbot-configs

your upper end could be changing very soon with the new additions, but yea rougly a 1/3 makes sense now

Comment 7

[Aki Sasaki (not active)]

Attachment: hm, uploaded wrong patch earlier (buildbotcustom)

Comment 8

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

Comment on attachment 559252 [details] [diff] [review]
hm, uploaded wrong patch earlier (buildbotcustom)

Review of attachment 559252 [details] [diff] [review]:
-----------------------------------------------------------------

Looks good.

Comment 9

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

Comment on attachment 559020 [details] [diff] [review]
split try and non-try tegra pools, buildbot-configs

Review of attachment 559020 [details] [diff] [review]:
-----------------------------------------------------------------

Guess I don't know much about how try works for mobile, but are no changes required in buildbot-configs/mozilla/*_config.py?

Looks good otherwise.

Comment 10

[Aki Sasaki (not active)]

(In reply to Chris Cooper [:coop] from comment #9)
> Comment on attachment 559020 [details] [diff] [review]
> split try and non-try tegra pools, buildbot-configs
> 
> Review of attachment 559020 [details] [diff] [review]:
> -----------------------------------------------------------------
> 
> Guess I don't know much about how try works for mobile, but are no changes
> required in buildbot-configs/mozilla/*_config.py?
> 
> Looks good otherwise.

I *think* since I'm only updating pools for the tegras inside misc.generateTalosBranchObjects(), I'm ok here.

Comment 11

[Aki Sasaki (not active)]

Comment on attachment 559020 [details] [diff] [review]
split try and non-try tegra pools, buildbot-configs

http://hg.mozilla.org/build/buildbot-configs/rev/b1174cccecbf

Comment 12

[Aki Sasaki (not active)]

Comment on attachment 559252 [details] [diff] [review]
hm, uploaded wrong patch earlier (buildbotcustom)

http://hg.mozilla.org/build/buildbotcustom/rev/7c9b75d464c2

Comment 13

[Armen [:armenzg]]

Merged to production and reconfigured.