590720 - update buildbot, puppet, opsi configs for new ix machines

Status: RESOLVED FIXED

(Release Engineering :: General, defect, P2)

Description

[[e-mail] (:bhearsum)]

Depends on bug 588960 because we need to know what their FQDNs are going to be.

Comment 1

[[e-mail] (:bhearsum)]

assigning to me for now

Comment 2

[[e-mail] (:bhearsum)]

Attachment: adjust buildbot-tac.py for new hostnames

This patch updates buildbot-tac.py to be compatible with the hostnames discussed in John's recent mail. We can assume anything with 'slave' in its name is a build slave, because test machines do not have that. With the proposed simplications in slave names, 'linux', 'darwin', and 'win' should be all we need to identify different platforms. As usual for staging slaves, they'll be sent to staging-master on first boot.

Comment 3

[Mike Taylor [:bear]]

Comment on attachment 469856 [details] [diff] [review]
adjust buildbot-tac.py for new hostnames

always a joy to see non-trivial if's collapse like that

r+ with the caveat that it lands only after bm-xserve-* renaming

Comment 4

[[e-mail] (:bhearsum)]

Attachment: add new ix machines to buildbot configs

This patch adds the new ix machines to buildbot configs. I've also dropped the long gone mv-moz2-linux-ix-slave24 from the staging configs.

The split is:
10 Linux to try, 7 Linux to production
20 Windows to try, 5 Windows to production

I went over wait times for the last week and Linux doesn't seem to be in too bad of shape on try or production most of the time. However, we need better burst capacity on try, so I put more there. For windows, the production wait times are actually quite good, but try is insufferable. Need most of them there.

Comment 5

[[e-mail] (:bhearsum)]

Attachment: add new linux ix machines to puppet manifests

These are headed to Castro to start with, so MPT needs to set them as "mv-build-node", and they need to be in mv-production.pp. We may need to add 10.250.52.0/24 to the fileserver configs, waiting to hear from jabba on that.

Comment 7

[[e-mail] (:bhearsum)]

Attachment: changeset: 58:2531531fddcf

Comment 8

[[e-mail] (:bhearsum)]

Attachment: updated patch to reserve some for staging

01 and 02 of both platforms are staging only in this version

Comment 9

[[e-mail] (:bhearsum)]

Comment on attachment 469884 [details] [diff] [review]
changeset:   58:2531531fddcf

changeset:   58:2531531fddcf

Comment 10

[[e-mail] (:bhearsum)]

Comment on attachment 469883 [details] [diff] [review]
add new linux ix machines to puppet manifests

changeset:   212:070907bc3f41

Comment 11

[[e-mail] (:bhearsum)]

Comment on attachment 469856 [details] [diff] [review]
adjust buildbot-tac.py for new hostnames

ac05929dc0b1

Comment 12

[[e-mail] (:bhearsum)]

Attachment: bump build tools puppet package version

We need to bump the version of the build tools repo on the ref machines. Here's the Puppet side, with a .spec part to come.

Comment 13

[[e-mail] (:bhearsum)]

Attachment: bump build-tools spec file

Comment 14

[Nick Thomas [:nthomas] (UTC+12)]

Landed
  http://hg.mozilla.org/build/puppet-manifests/rev/070907bc3f41
to fix mv-puppet-production and bug 591803.

Comment 15

[[e-mail] (:bhearsum)]

Comment on attachment 469996 [details] [diff] [review]
bump build tools puppet package version

changeset:   214:79cb19ba1ff0

Comment 16

[[e-mail] (:bhearsum)]

build-tools package update landed successfully (affects all linux/mac build machines, I tested one of each OS).

Comment 17

[[e-mail] (:bhearsum)]

Will need to land the configs pretty soon now.

Comment 18

[[e-mail] (:bhearsum)]

Comment on attachment 469889 [details] [diff] [review]
updated patch to reserve some for staging

Landed the staging hunks: changeset:   2946:a60fc822636c

Comment 19

[[e-mail] (:bhearsum)]

Comment on attachment 469889 [details] [diff] [review]
updated patch to reserve some for staging

Turns out the rest of this patch is invalid, win32-ix-slaveNN is too many characters for Windows :(.

Comment 20

[[e-mail] (:bhearsum)]

Attachment: fix slavenames in buildbot-configs, staging only

Comment 21

[[e-mail] (:bhearsum)]

Attachment: fix slavenames in opsi configs

Comment 22

[[e-mail] (:bhearsum)]

Attachment: let buildbot-tac.py handle new slavenames

Comment 23

[[e-mail] (:bhearsum)]

Comment on attachment 471996 [details] [diff] [review]
fix slavenames in opsi configs

changeset:   62:9afce7d8597e

Comment 24

[[e-mail] (:bhearsum)]

Comment on attachment 471997 [details] [diff] [review]
let buildbot-tac.py handle new slavenames

changeset:   851:758886cfca08

Comment 25

[[e-mail] (:bhearsum)]

Comment on attachment 471995 [details] [diff] [review]
fix slavenames in buildbot-configs, staging only

changeset:   2951:3299424c9539

Comment 26

[[e-mail] (:bhearsum)]

Attachment: updated patch for production ix buildbot configs

Comment 27

[Nick Thomas [:nthomas] (UTC+12)]

Attachment: Fix up staging puppet for linux

Looks like a copy-and-paste-o.

Comment 28

[[e-mail] (:bhearsum)]

Comment on attachment 473400 [details] [diff] [review]
Fix up staging puppet for linux

Thanks for catching this

Comment 29

[[e-mail] (:bhearsum)]

Comment on attachment 473400 [details] [diff] [review]
Fix up staging puppet for linux

changeset:   219:735c41db521c

I updated the checkout on staging-puppet.

Comment 30

[Nick Thomas [:nthomas] (UTC+12)]

Attachment: Add linux-ix-slave{16,17} to puppet

These two are up but not connecting to puppet.

Comment 31

[Nick Thomas [:nthomas] (UTC+12)]

Comment on attachment 473793 [details] [diff] [review]
Add linux-ix-slave{16,17} to puppet

http://hg.mozilla.org/build/puppet-manifests/rev/b11560284a45
mv-p-p updated, slaves rebooted to sync up.

Comment 32

[[e-mail] (:bhearsum)]

These machines have started making their way to production in bug 595274. I think we're all done here.