Page MenuHomePhabricator

Update LPL cloud VPS instances to Trixie
Closed, ResolvedPublic8 Estimated Story Points

Description

Debian's latest release 13 (Trixie) is now available in the cloud VPS.

Upgrade the LPL test instances from debian-12.0-bookworm to the latest Debian release.

Upgrade StatusInstancessh name Notes
Donelpl-cx-sxlpl-cx-sx.language.eqiad1.wikimedia.cloudContent and Section Translation (auto update)
Donelpl-serviceslpl-services.language.eqiad1.wikimedia.cloudcxserver and Apertium services
Donelpl-mleb-stablelpl-mleb-stable.language.eqiad1.wikimedia.cloudMLEB Stable version
Donelpl-mleb-masterlpl-mleb-master.language.eqiad1.wikimedia.cloudMLEB Master version (auto update)
Donelpl-recommendlpl-recommend.language.eqiad1.wikimedia.cloudRecommendation API
See: T412340mintmint.language.eqiad1.wikimedia.cloudMinT service

Event Timeline

KartikMistry changed the task status from Open to In Progress.Aug 26 2025, 8:08 AM
KartikMistry moved this task from Backlog to In Progress on the LPL Essential (2025 Jul-Oct) board.
A_smart_kitten renamed this task from Udpate LPL cloud VPS instance to Trixie to Update LPL cloud VPS instance to Trixie.Aug 26 2025, 8:10 AM
Nikerabbit renamed this task from Update LPL cloud VPS instance to Trixie to Update LPL cloud VPS instances to Trixie.Sep 1 2025, 8:17 AM
Nikerabbit updated the task description. (Show Details)
Nikerabbit triaged this task as Medium priority.Sep 1 2025, 8:43 AM
Nikerabbit set the point value for this task to 8.Sep 1 2025, 1:40 PM

We are running out of CPU/RAM quota, so I'll go through replacing instances one by one.

Holding on updating lpl-services till we are done with T394982

There are some issues with DNS making MLEB accessible using subdomains like stable.lpl-mleb.wmcloud.org. I'll take approach that's better for the long term.

Both MLEB instances currently fatal out when loading Special:Version:

[b603bde6e0f54d5370a502d9] /index.php/Special:Version Error: Class "Elastica\Client" not found

Backtrace:

from /var/www/core/extensions/Elastica/includes/ElasticaConnection.php(97)
#0 /var/www/core/extensions/Elastica/includes/ElasticaConnection.php(68): MediaWiki\Extension\Elastica\ElasticaConnection->getClient()
#1 /var/www/core/extensions/CirrusSearch/includes/Connection.php(130): MediaWiki\Extension\Elastica\ElasticaConnection->setConnectTimeout()
#2 /var/www/core/extensions/CirrusSearch/includes/Hooks.php(535): CirrusSearch\Connection->__construct()

See: https://lpl-mleb-stable.wmcloud.org/index.php/Special:Version

Both MLEB instances currently fatal out when loading Special:Version:

[b603bde6e0f54d5370a502d9] /index.php/Special:Version Error: Class "Elastica\Client" not found

Backtrace:

from /var/www/core/extensions/Elastica/includes/ElasticaConnection.php(97)
#0 /var/www/core/extensions/Elastica/includes/ElasticaConnection.php(68): MediaWiki\Extension\Elastica\ElasticaConnection->getClient()
#1 /var/www/core/extensions/CirrusSearch/includes/Connection.php(130): MediaWiki\Extension\Elastica\ElasticaConnection->setConnectTimeout()
#2 /var/www/core/extensions/CirrusSearch/includes/Hooks.php(535): CirrusSearch\Connection->__construct()

See: https://lpl-mleb-stable.wmcloud.org/index.php/Special:Version

Fixed. Thanks!

I'm unable to SSH to the instance, maybe an issue on my end.

Translation memory still appears unusable.

Please do the following:

# 1. Delete all Elasticsearch index
curl -X DELETE 'http://localhost:9200/_all'

# 2. Run ttmserver-export.php (extensions/Translate/scripts/ttmserver-export.php)

Note that new instances are lpl-mleb-master2.language.eqiad1.wikimedia.cloud and lpl-mleb-stable2.language.eqiad1.wikimedia.cloud

Both MLEB instances are unstable with low RAM. I'll be recreating them.

Both MLEB instances are unstable with low RAM. I'll be recreating them.

Both MLEB instances are up and running now.

MediaWiki Job runner is enabled in both MLEB instances.

MediaWiki Job runner is enabled in both MLEB instances.

Yup, can confirm that it works fine now.

Note that new instances are lpl-mleb-master2.language.eqiad1.wikimedia.cloud and lpl-mleb-stable2.language.eqiad1.wikimedia.cloud

I tried using these from bastion.wmcloud.org:

$ ssh lpl-mleb-master2.language.eqiad1.wikimedia.cloud
ssh: Could not resolve hostname lpl-mleb-master2.language.eqiad1.wikimedia.cloud: Name or service not known

Any ideas what could be wrong?

$ ssh lpl-mleb-master2.language.eqiad1.wikimedia.cloud
ssh: Could not resolve hostname lpl-mleb-master2.language.eqiad1.wikimedia.cloud: Name or service not known

Any ideas what could be wrong?

MLEB Stable: lpl-mleb-stable.language.eqiad1.wikimedia.cloud
MLEB Master: lpl-mleb-master.language.eqiad1.wikimedia.cloud

I'm also investigating the MinT instance's frequent downtime, so recreating the instance creation is taking longer than usual. I'll update my findings here.

KartikMistry changed the point value for this task from 8 to 6.
KartikMistry changed the status of subtask T412340: Update MinT cloud VPS instances to Trixie from Open to In Progress.
KartikMistry moved this task from In Progress to Done on the LPL Essential (FY2025-26 Q2) board.
Nikerabbit changed the point value for this task from 6 to 8.
Nikerabbit subscribed.

6 is not part of our estimation scale, should be either 4 or 8.