Closed Bug 864741 Opened 12 years ago Closed 12 years ago

After we disable talos jobs (partially) on Fedora minis we should re-purpose some of them

Categories

(Infrastructure & Operations Graveyard :: CIDuty, task, P2)

x86
Linux

Tracking

(Not tracked)

RESOLVED FIXED

People

(Reporter: armenzg, Assigned: armenzg)

References

Details

(Whiteboard: [re-js][re-nad-20130501])

Attachments

(10 files)

1.28 KB, patch
Callek
: review+
armenzg
: checked-in+
Details | Diff | Splinter Review
1.85 KB, patch
Callek
: review+
armenzg
: checked-in+
Details | Diff | Splinter Review
961 bytes, patch
rail
: review+
armenzg
: checked-in+
Details | Diff | Splinter Review
867 bytes, patch
Callek
: review+
armenzg
: checked-in+
Details | Diff | Splinter Review
578 bytes, patch
Callek
: review+
armenzg
: checked-in+
Details | Diff | Splinter Review
544 bytes, patch
rail
: review+
armenzg
: checked-in+
Details | Diff | Splinter Review
699 bytes, patch
rail
: review+
armenzg
: checked-in+
Details | Diff | Splinter Review
787 bytes, patch
rail
: review+
armenzg
: checked-in+
Details | Diff | Splinter Review
1.25 KB, patch
rail
: review+
armenzg
: checked-in+
Details | Diff | Splinter Review
1.59 KB, patch
rail
: review+
armenzg
: checked-in+
Details | Diff | Splinter Review
In bug 863903 we're disabling talos on FF23 based project branches. After a couple of days I will be able to determine how many we can re-purpose. This bug is to track the work after that bug.
Summary: After we disable talos jobs (partially) on Fedora minis we should re-purpose them → After we disable talos jobs (partially) on Fedora minis we should re-purpose some of them
Blocks: 864866
Whiteboard: [re-js]
Assignee: nobody → armenzg
update: next action date: 2013-05-01 - 04-25 - fedora talos jobs are expected to be disabled today (Apr 25). - 05-01 - we'll be in a position to identify the number of machines to be reimaged. Likely in the 30 machine range.
Whiteboard: [re-js] → [re-js][re-nad-20130501]
We disabled talos jobs late in the day yesterday (2013-04-25 18:56:31 PDT) so the wait time numbers for yesterday do not help. [1] Nevertheless, if I tweak this URL I can see how we did in the last 12 hours [2]: https://secure.pub.build.mozilla.org/buildapi/reports/waittimes/testpool?starttime=1366938000&endtime=1366992000 and I can compare to the same time period for a day before: [3] https://secure.pub.build.mozilla.org/buildapi/reports/waittimes/testpool?starttime=1366851600&endtime=1366905600 I'm going to disable 15 of our fed64 machines and 15 of the fed32 machines and see on Monday how we do today. talos-r3-fed-001 -> 002 is loaned talos-r3-fed-003 talos-r3-fed-004 talos-r3-fed-005 talos-r3-fed-006 talos-r3-fed-007 talos-r3-fed-008 talos-r3-fed-009 talos-r3-fed-010 talos-r3-fed-011 talos-r3-fed-012 talos-r3-fed-013 talos-r3-fed-014 talos-r3-fed-015 talos-r3-fed-016 talos-r3-fed64-031 talos-r3-fed64-032 talos-r3-fed64-033 talos-r3-fed64-034 --> not existent talos-r3-fed64-036 talos-r3-fed64-037 talos-r3-fed64-038 talos-r3-fed64-039 talos-r3-fed64-040 talos-r3-fed64-041 talos-r3-fed64-042 talos-r3-fed64-043 talos-r3-fed64-044 talos-r3-fed64-045 talos-r3-fed64-046 [1] fedora: 3592 0: 3464 96.44% 15: 92 2.56% 30: 35 0.97% 45: 1 0.03% fedora64: 874 0: 857 98.05% 15: 17 1.95% ##################################### [2] fedora: 1297 0 1297 100.00% fedora64: 291 0 291 100.00% ##################################### [3] fedora: 1789 0 1789 100.00% fedora64: 622 0 617 99.20% 15 4 0.64% 30 1 0.16%
Priority: -- → P2
Blocks: 866194
I removed too many from each pool :( I will put 10 machines back for each pool. Friday's wait times: #################### fedora: 3814 86.37% fedora64: 1001 54.65% The list for now will be this (I've added a new machine talos-r3-fed-002): | talos-r3-fed-001 | | talos-r3-fed-002 | | talos-r3-fed-003 | | talos-r3-fed-004 | | talos-r3-fed-005 | | talos-r3-fed64-031 | | talos-r3-fed64-032 | | talos-r3-fed64-033 | | talos-r3-fed64-034 | | talos-r3-fed64-036 |
Attachment #743094 - Flags: review?(bugspam.Callek)
Attachment #743095 - Flags: review?(bugspam.Callek)
Attachment #743096 - Flags: review?(bugspam.Callek)
Attachment #743093 - Flags: review?(bugspam.Callek) → review+
Attachment #743094 - Flags: review?(bugspam.Callek) → review+
Attachment #743095 - Flags: review?(bugspam.Callek) → review+
Attachment #743096 - Flags: review?(bugspam.Callek) → review+
I forgot this patch.
Attachment #743166 - Flags: review?(bugspam.Callek)
Comment on attachment 743166 [details] [diff] [review] [opsi] add winxp machines Because I like you I'll let you land this, but I will never admit to reviewing an opsi patch, ever.
Attachment #743166 - Flags: review?(bugspam.Callek) → review+
Attachment #743093 - Flags: checked-in+
Attachment #743094 - Flags: checked-in+
Attachment #743095 - Flags: checked-in+
Attachment #743096 - Flags: checked-in+
Attachment #743166 - Flags: checked-in+
Today we had better wait times (we didn't have as high of a load though - only 42K jobs). I will disable another 5 from each and see how we do tomorrow. mysql> select name from slaves where notes like '%864741%' and envid=2 order by name; +--------------------+ | name | +--------------------+ | talos-r3-fed-006 | | talos-r3-fed-007 | | talos-r3-fed-008 | | talos-r3-fed-009 | | talos-r3-fed-010 | | talos-r3-fed64-037 | | talos-r3-fed64-038 | | talos-r3-fed64-039 | | talos-r3-fed64-040 | | talos-r3-fed64-041 | +--------------------+ 10 rows in set (0.11 sec) Monday's wait times: #################### fedora: 3346 - 99.13% fedora64: 857 - 98.95% Friday's wait times: #################### fedora: 3814 86.37% fedora64: 1001 54.65%
No longer blocks: 866194
Depends on: 866194
The following WinXP machines have been added to the production pool: talos-r3-xp-128 talos-r3-xp-129 -- talos-r3-xp-131 Waiting on imaging for the remaining ones.
I'm putting 3 fed64 machines back so we won't score so lowly on a high load day. We won't hit perfect with 3 more but we're OK to have better XP wait times than Fed64 as the end-to-end times are worst on Windows. This means that I will request 7 more machines to be re-imaged: | talos-r3-fed-006 | | talos-r3-fed-007 | | talos-r3-fed-008 | | talos-r3-fed-009 | | talos-r3-fed-010 | | talos-r3-fed64-037 | | talos-r3-fed64-038 | | talos-r3-fed64-039 | Tuesday's wait times (-10 machines): #################################### Total jobs: 49k fedora: 3883 - 93.79% fedora64: 1112 - 76.35% Monday's wait times (-5 machines): ################################## Total jobs: 42k fedora: 3346 - 99.13% fedora64: 857 - 98.95% Friday's wait times (-15 machines): ################################### Total jobs: 39k fedora: 3814 86.37% fedora64: 1001 54.65%
Attachment #745180 - Flags: review?
Attachment #745181 - Flags: review?
Attachment #745183 - Flags: review?
Attachment #745182 - Flags: review?
The list of machines.
Attachment #745186 - Flags: review?(rail)
Attachment #745183 - Flags: review? → review+
Attachment #745182 - Flags: review? → review+
Attachment #745186 - Flags: review?(rail) → review+
Attachment #745181 - Flags: review? → review+
Attachment #745180 - Flags: review? → review+
Attachment #745180 - Flags: checked-in+
Attachment #745181 - Flags: checked-in+
Attachment #745182 - Flags: checked-in+
Attachment #745183 - Flags: checked-in+
Attachment #745186 - Flags: checked-in+
In production.
All the machines that have been set aside have been re-imaged. WinXP wait times are in the 90+% close to 95%.
Status: NEW → RESOLVED
Closed: 12 years ago
Resolution: --- → FIXED
Product: mozilla.org → Release Engineering
Product: Release Engineering → Infrastructure & Operations
Product: Infrastructure & Operations → Infrastructure & Operations Graveyard
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: