Closed Bug 802589 Opened 12 years ago Closed 12 years ago

Reduce underutilized Leopard pool

Categories

(Infrastructure & Operations Graveyard :: CIDuty, task, P3)

x86
macOS

Tracking

(Not tracked)

RESOLVED FIXED

People

(Reporter: armenzg, Assigned: armenzg)

References

Details

Attachments

(4 files, 1 obsolete file)

> We got 22 production Leopard slaves if I can count correctly.
> Yesterday we had only 56 Leopard jobs.
> If I divide that between 10 hours of the day comes to 5.6 jobs/hour.
>
> Any inconvenience if we go down to 15 Leopard slaves? I think we could go even
> lower but I want to accomodate for failing slaves and even higher loads.
> If you think we could go lower please let me know. I would rather see wait
> times on Leopard test jobs.

No one has raised objections to this plan.

I will disable a bunch of them and evaluate after a week or so.
I noticed something interesting.
We run jetpack jobs for Leopard when we should not.
I will attach a patch to address that as well.

I have disabled a bunch of slaves and left 15 production ones and 1 staging up and running.
Priority: -- → P3
Jetpack is bug 778969.
On 10-19:
leopard: 88
  0:       83    94.32%
 15:        5     5.68%

On 10-20:
leopard: 17
  0:       17   100.00%

It seems that bug 778969 took a lot of the load.
I have disabled down to 7 production slaves and 1 staging slave.

We would be able to recuperate 17 rev3 minis.
On 10-22:
leopard: 115
  0:       57    49.57%
 15:       39    33.91%
 30:       15    13.04%
 45:        2     1.74%
 60:        1     0.87%
 75:        0     0.00%
 90+:        1     0.87%

I have re-enabled 4 slaves back.

Our wait times would depend on how many changes land for:
* TB comm-esr10
* FF mozilla-esr10
* TB comm-release
* FF mozilla-release

I guess that as soon as version 17 goes into release and the esr-17 becomes the new esr we will be able to completely eliminate the pool.
People keep on pushing.

Adding 1 more slave back and if we have wait times some days for the next 4 weeks let it be.

leopard: 163
  0:      106    65.03%
 15:       35    21.47%
 30:       12     7.36%
 45:        6     3.68%
 60:        3     1.84%
 75:        0     0.00%
 90+:        1     0.61%
Depends on: 805008
These need to be added:
talos-r3-leopard-002   talos-r3-xp-096
talos-r3-leopard-010   talos-r3-xp-097
talos-r3-leopard-017   talos-r3-xp-098
talos-r3-leopard-018   talos-r3-xp-099
talos-r3-leopard-019   talos-r3-xp-100

talos-r3-leopard-020   talos-r3-w7-100
talos-r3-leopard-021   talos-r3-w7-101
talos-r3-leopard-022   talos-r3-w7-102
talos-r3-leopard-023   talos-r3-w7-103
talos-r3-leopard-024   talos-r3-w7-104


mysql> select * from slaves where name like 'talos-r3-xp-095';
+---------+-----------------+----------+--------+---------+-----------+------+---------+-------+--------+----------------+-----------------+---------+------------------+-------+--------------+
| slaveid | name            | distroid | bitsid | speedid | purposeid | dcid | trustid | envid | poolid | basedir        | locked_masterid | enabled | current_masterid | notes | custom_tplid |
+---------+-----------------+----------+--------+---------+-----------+------+---------+-------+--------+----------------+-----------------+---------+------------------+-------+--------------+
|    2358 | talos-r3-xp-095 |        7 |      1 |       1 |         4 |    7 |       4 |     2 |     19 | C:\talos-slave |            NULL |       1 |               61 |       |         NULL |
+---------+-----------------+----------+--------+---------+-----------+------+---------+-------+--------+----------------+-----------------+---------+------------------+-------+--------------+

mysql> select * from slaves where name like 'talos-r3-w7-095';
+---------+-----------------+----------+--------+---------+-----------+------+---------+-------+--------+----------------+-----------------+---------+------------------+-------+--------------+
| slaveid | name            | distroid | bitsid | speedid | purposeid | dcid | trustid | envid | poolid | basedir        | locked_masterid | enabled | current_masterid | notes | custom_tplid |
+---------+-----------------+----------+--------+---------+-----------+------+---------+-------+--------+----------------+-----------------+---------+------------------+-------+--------------+
|    2374 | talos-r3-w7-095 |        8 |      1 |       1 |         4 |    7 |       4 |     2 |     19 | C:\talos-slave |            NULL |       1 |               61 | NULL  |         NULL |
+---------+-----------------+----------+--------+---------+-----------+------+---------+-------+--------+----------------+-----------------+---------+------------------+-------+--------------+
Attached patch buildbot-configsSplinter Review
Attachment #677018 - Flags: review?(coop)
Attached patch opsiSplinter Review
Attachment #677019 - Flags: review?(coop)
Attached patch slavealloc (obsolete) — Splinter Review
Attachment #677020 - Flags: review?(coop)
Attached patch graphsdbSplinter Review
Attachment #677021 - Flags: review?(coop)
Attachment #677018 - Flags: review?(coop) → review+
Attachment #677019 - Flags: review?(coop) → review+
Attachment #677020 - Flags: review?(coop) → review+
Attachment #677021 - Flags: review?(coop) → review+
Depends on: 807450
I updated the documentation to have less chances of doing it wrong.
Attachment #677020 - Attachment is obsolete: true
The buildbot-configs patch is in production.
xp 097 was down and is now back up (bug 807320). Couldn't get it into the production pool because it's not in OPSI yet.
Depends on: 808421
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: