Page MenuHomePhabricator

Turn down mwmaint production servers
Closed, ResolvedPublic

Description

  • Copy homedirs to /srv/homedirs/mwmaint{1002,2002} on deploy1003
  • Communicate upcoming turn down to wikitech-l
  • Restrict access to ops group members (soft turndown)
  • Decommission mwmaint servers

Timeline TBD

Event Timeline

Clement_Goubert triaged this task as Medium priority.

Change #1163731 had a related patch set uploaded (by Clément Goubert; author: Clément Goubert):

[operations/puppet@production] P::mediawiki::maintenance: rsync to deployment server

https://gerrit.wikimedia.org/r/1163731

Change #1163731 merged by Clément Goubert:

[operations/puppet@production] P::mediawiki::maintenance: rsync to deployment server

https://gerrit.wikimedia.org/r/1163731

hey, before we plug the switch, would it be possible to adjust the default mysql prompt on the deployment machine? I like mwmaint tells me where am I connected, while deploy* doesn't do that.

Change #1164139 had a related patch set uploaded (by Clément Goubert; author: Clément Goubert):

[operations/puppet@production] O:kubernetes::deployment_server: mariadb client

https://gerrit.wikimedia.org/r/1164139

Change #1164139 merged by Clément Goubert:

[operations/puppet@production] O:kubernetes::deployment_server: mariadb client

https://gerrit.wikimedia.org/r/1164139

hey, before we plug the switch, would it be possible to adjust the default mysql prompt on the deployment machine? I like mwmaint tells me where am I connected, while deploy* doesn't do that.

Did adding profile::mariadb::client to deployment servers fix your prompt issue?

hey, before we plug the switch, would it be possible to adjust the default mysql prompt on the deployment machine? I like mwmaint tells me where am I connected, while deploy* doesn't do that.

Did adding profile::mariadb::client to deployment servers fix your prompt issue?

Yes, thank you!

Change #1167195 had a related patch set uploaded (by Clément Goubert; author: Clément Goubert):

[operations/puppet@production] check_user: Use deploy instead of mwmaint

https://gerrit.wikimedia.org/r/1167195

Change #1167196 had a related patch set uploaded (by Clément Goubert; author: Clément Goubert):

[operations/puppet@production] mwaint: Remove from scap

https://gerrit.wikimedia.org/r/1167196

Change #1167197 had a related patch set uploaded (by Clément Goubert; author: Clément Goubert):

[operations/puppet@production] mwmaint: deprecate mwmaint servers

https://gerrit.wikimedia.org/r/1167197

Change #1167195 abandoned by Clément Goubert:

[operations/puppet@production] check_user: Use deploy instead of mwmaint

Reason:

Script is broken and will likely be removed

https://gerrit.wikimedia.org/r/1167195

Change #1167196 merged by Clément Goubert:

[operations/puppet@production] mwaint: Remove from scap

https://gerrit.wikimedia.org/r/1167196

Mentioned in SAL (#wikimedia-operations) [2025-07-09T10:14:25Z] <claime> Cutting off access to mwmaint servers - T397017

Change #1167197 merged by Clément Goubert:

[operations/puppet@production] mwmaint: deprecate mwmaint servers

https://gerrit.wikimedia.org/r/1167197

Change #1171983 had a related patch set uploaded (by Clément Goubert; author: Clément Goubert):

[operations/dns@master] wmnet: Remove maintenance.eqiad.wmnet record

https://gerrit.wikimedia.org/r/1171983

Change #1171983 merged by Clément Goubert:

[operations/dns@master] wmnet: Remove maintenance.eqiad.wmnet record

https://gerrit.wikimedia.org/r/1171983

As part of deploying the train, deploy1003 had a disk space alarm for /srv:

CRITICAL: DISK CRITICAL - free space: /srv 8607 MB (3% inode=69%)
df -h /srv
Filesystem           Size  Used Avail Use% Mounted on
/dev/mapper/vg0-srv  277G  255G  8.3G  97% /srv
du -d2 -m -x /srv 2>/dev/null |sort -rn|head -n10
101093	/srv
37541	/srv/mediawiki-staging
36861	/srv/deployment
26615	/srv/homedirs
13592	/srv/homedirs/mwmaint1002
13024	/srv/homedirs/mwmaint2002
10120	/srv/mediawiki-staging/php-1.45.0-wmf.12
10087	/srv/mediawiki-staging/php-1.45.0-wmf.11
10021	/srv/mediawiki-staging/php-1.45.0-wmf.13
7917	/srv/deployment/analytics

The homedirs consume 26G or 10% of the available disk space. That is not that much in the grand scheme of things :)

I plan on removing the homedirs soon-ish, but I will add 30GB to the vg to negate the issue in the meantime.

Clement_Goubert updated the task description. (Show Details)
root@deploy1003:~# lvextend -L+30G /dev/vg0/srv
  Size of logical volume vg0/srv changed from <281.92 GiB (72171 extents) to <311.92 GiB (79851 extents).
  Logical volume vg0/srv successfully resized.
root@deploy1003:~# resize2fs /dev/vg0/srv
resize2fs 1.46.2 (28-Feb-2021)
Filesystem at /dev/vg0/srv is mounted on /srv; on-line resizing required
old_desc_blocks = 36, new_desc_blocks = 39
The filesystem on /dev/vg0/srv is now 81767424 (4k) blocks long.

root@deploy1003:~# df -h /srv
Filesystem           Size  Used Avail Use% Mounted on
/dev/mapper/vg0-srv  306G  255G   37G  88% /srv

I am happy to see there was plenty of disk space on the volume group \o/ Thank you!

Change #1190309 had a related patch set uploaded (by Jasmine; author: Jasmine):

[operations/dns@master] wmnet: remove mwmaint discovery aliases since turning down production servers [0]

https://gerrit.wikimedia.org/r/1190309

Change #1190309 merged by Jasmine:

[operations/dns@master] wmnet: remove mwmaint discovery aliases since turning down production servers [0]

https://gerrit.wikimedia.org/r/1190309

Change #1219189 had a related patch set uploaded (by Krinkle; author: Krinkle):

[operations/puppet@production] scap: Remove unused mwmaint config, obsolete wikitech/php7 comments

https://gerrit.wikimedia.org/r/1219189

Change #1219189 merged by RLazarus:

[operations/puppet@production] scap: Remove unused mwmaint config, obsolete wikitech/php7 comments

https://gerrit.wikimedia.org/r/1219189

Change #1228454 had a related patch set uploaded (by Muehlenhoff; author: Muehlenhoff):

[operations/puppet@production] Remove ferm/rsync/tcpircbot settings

https://gerrit.wikimedia.org/r/1228454

Change #1228457 had a related patch set uploaded (by Muehlenhoff; author: Muehlenhoff):

[operations/puppet@production] Fix the description of what restricted does after mwmaint* decom

https://gerrit.wikimedia.org/r/1228457

Change #1228457 merged by Muehlenhoff:

[operations/puppet@production] Fix the description of what restricted does after mwmaint* decom

https://gerrit.wikimedia.org/r/1228457

Change #1228454 merged by Muehlenhoff:

[operations/puppet@production] Remove ferm/rsync/tcpircbot settings

https://gerrit.wikimedia.org/r/1228454

Change #1228507 had a related patch set uploaded (by Marostegui; author: Marostegui):

[operations/puppet@production] production-m5.sql.erb: Remove old mwmaint IP

https://gerrit.wikimedia.org/r/1228507

Change #1228507 merged by Marostegui:

[operations/puppet@production] production-m5.sql.erb: Remove old mwmaint IP

https://gerrit.wikimedia.org/r/1228507