After repurposing mw-parsoid in T386246, we can retire the LVS service as it is not used anymore.
Description
Details
Related Objects
Event Timeline
@jijiki - I've optimistically triaged this to Medium and moved it to Scheduled, since my understanding is that you'll be working on this soon. Please adjust if needed. Thank you!
Change #1261428 had a related patch set uploaded (by Effie Mouzeli; author: Effie Mouzeli):
[operations/dns@master] mw-parsoid: remove DNS of mw-parsoid LVS service 1
Change #1261430 had a related patch set uploaded (by Effie Mouzeli; author: Effie Mouzeli):
[operations/puppet@production] mw-parsoid: Remove probes 2
Change #1261433 had a related patch set uploaded (by Effie Mouzeli; author: Effie Mouzeli):
[operations/puppet@production] mw-parsoid: switch to service_setup 3
Change #1262025 had a related patch set uploaded (by Effie Mouzeli; author: Effie Mouzeli):
[operations/dns@master] mw-parsoid: add CNAMES
Change #1262026 had a related patch set uploaded (by Effie Mouzeli; author: Effie Mouzeli):
[operations/puppet@production] trafficserver: update mw-parsoid XWD entries
Change #1262052 had a related patch set uploaded (by Effie Mouzeli; author: Effie Mouzeli):
[operations/puppet@production] mw-parsoid: remove service definition 4
Change #1262054 had a related patch set uploaded (by Effie Mouzeli; author: Effie Mouzeli):
[operations/puppet@production] envoy: remove mw-parsoid listener
Change #1261428 merged by BCornwall:
[operations/dns@master] mw-parsoid: remove DNS of mw-parsoid LVS service 1
Change #1261430 merged by BCornwall:
[operations/puppet@production] mw-parsoid: Remove probes 2
Change #1261433 merged by BCornwall:
[operations/puppet@production] mw-parsoid: switch to service_setup 3
Mentioned in SAL (#wikimedia-operations) [2026-03-30T19:05:59Z] <brett@cumin2002> START - Cookbook sre.loadbalancer.restart-pybal rolling-restart of pybal on A:lvs-secondary-codfw and A:lvs (T420468)
Mentioned in SAL (#wikimedia-operations) [2026-03-30T19:06:30Z] <brett@cumin2002> END (PASS) - Cookbook sre.loadbalancer.restart-pybal (exit_code=0) rolling-restart of pybal on A:lvs-secondary-codfw and A:lvs (T420468)
Mentioned in SAL (#wikimedia-operations) [2026-03-30T19:08:06Z] <brett@cumin2002> START - Cookbook sre.loadbalancer.restart-pybal rolling-restart of pybal on A:lvs-secondary-eqiad and A:lvs (T420468)
Mentioned in SAL (#wikimedia-operations) [2026-03-30T19:10:32Z] <brett@cumin2002> END (PASS) - Cookbook sre.loadbalancer.restart-pybal (exit_code=0) rolling-restart of pybal on A:lvs-secondary-eqiad and A:lvs (T420468)
Mentioned in SAL (#wikimedia-operations) [2026-03-30T19:11:28Z] <brett@cumin2002> START - Cookbook sre.loadbalancer.restart-pybal rolling-restart of pybal on A:lvs-low-traffic-codfw and A:lvs (T420468)
Mentioned in SAL (#wikimedia-operations) [2026-03-30T19:11:59Z] <brett@cumin2002> END (PASS) - Cookbook sre.loadbalancer.restart-pybal (exit_code=0) rolling-restart of pybal on A:lvs-low-traffic-codfw and A:lvs (T420468)
Mentioned in SAL (#wikimedia-operations) [2026-03-30T19:12:15Z] <brett@cumin2002> START - Cookbook sre.loadbalancer.restart-pybal rolling-restart of pybal on A:lvs-low-traffic-eqiad and A:lvs (T420468)
Mentioned in SAL (#wikimedia-operations) [2026-03-30T19:12:57Z] <brett@cumin2002> END (PASS) - Cookbook sre.loadbalancer.restart-pybal (exit_code=0) rolling-restart of pybal on A:lvs-low-traffic-eqiad and A:lvs (T420468)
Mentioned in SAL (#wikimedia-operations) [2026-03-30T19:15:15Z] <brett> Delete mw-parsoid lvs service via sudo -i cumin A:lvs-low-traffic-codfw 'ipvsadm --delete-service --tcp-service 10.2.1.92:4452' - T420468
Mentioned in SAL (#wikimedia-operations) [2026-03-30T19:16:18Z] <brett> Delete mw-parsoid lvs service via sudo -i cumin A:lvs-secondary-codfw 'ipvsadm --delete-service --tcp-service 10.2.1.92:4452' - T420468
Mentioned in SAL (#wikimedia-operations) [2026-03-30T19:18:09Z] <brett> Delete mw-parsoid lvs service via sudo -i cumin A:lvs-secondary-eqiad 'ipvsadm --delete-service --tcp-service 10.2.2.92:4452' - T420468
Mentioned in SAL (#wikimedia-operations) [2026-03-30T19:18:18Z] <brett> Delete mw-parsoid lvs service via sudo -i cumin A:lvs-low-traffic-eqiad 'ipvsadm --delete-service --tcp-service 10.2.2.92:4452' - T420468
Change #1262052 merged by BCornwall:
[operations/puppet@production] mw-parsoid: remove service definition 4
Change #1262025 merged by Effie Mouzeli:
[operations/dns@master] mw-parsoid: add CNAMES
Change #1262026 merged by Effie Mouzeli:
[operations/puppet@production] trafficserver: update mw-parsoid XWD entries
Change #1265420 had a related patch set uploaded (by Elukey; author: Elukey):
[operations/puppet@production] profile::service_proxy::envoy: remove mw-parsoid
Change #1262054 merged by Brouberol:
[operations/puppet@production] envoy: remove mw-parsoid listener
Puppet on RESTBase is failing with:
Error: Could not retrieve catalog from remote server: Error 500 on SERVER: Server Error: Evaluation Error: Error while evaluating a Function Call, One and only one listener with name 'mw-parsoid' is expected (file: /srv/puppet_code/environments/production/modules/wmflib/functions/service/get_url.pp, line: 77, column: 13) on node restbase1031.eqiad.wmnet Warning: Not using cache on failed catalog Error: Could not retrieve catalog; skipping run
In this case 'One and only one...' is because there are zero listeners, this happened after the merging of r1262054. Allegedly we're not using RESTBase's parsoid endpoints for anything any more (/cc @daniel ?), but I don't know what it will do with a url that doesn't work at all (refuse to start? logspam? nothing at all?). Spoiler alert: I think we are about to find out.
Change #1265450 had a related patch set uploaded (by Eevans; author: Eevans):
[operations/puppet@production] restbase: unbreak puppet; address missing mw-parsoid listener
Change #1265452 had a related patch set uploaded (by Brouberol; author: Brouberol):
[operations/deployment-charts@master] fixtures: remove mw-parsoid listener from fixtures after it's been dropped
Change #1265452 merged by Brouberol:
[operations/deployment-charts@master] fixtures: remove mw-parsoid listener from fixtures after it's been dropped
Change #1265450 merged by Eevans:
[operations/puppet@production] restbase: address missing mw-parsoid listener (unbreak puppet)
r1265450 (merged) updated mw-parsoid to mw-api-int instead, and puppet is now running successfully. The transform endpoint does not work anymore though...
/en.wikipedia.org/v1/page/html/{title} (Get html by title from storage) is CRITICAL: Test Get html by title from storage returned the unexpected status 504 (expecting: 200); /en.wikipedia.org/v1/transform/wikitext/to/html/{title} (Transform wikitext to html) is CRITICAL: Test Transform wikitext to html returned the unexpected status 504 (expecting: 200)Because there is no local listener present...
eevans@restbase1031:~$ sudo grep parsoid_uri /etc/restbase/config-vars.yaml parsoid_uri: http://localhost:6501/w/rest.php eevans@restbase1031:~$ sudo ss -npl |grep 6501 eevans@restbase1031:~$
This results in some ECONNREFUSED errors...
... but as far as I can tell, we're only emitting errors for service checks (i.e. when check-restbase is invoked), so I guess this is "OK"?
mw-parsoid is still referenced as discovery listener in multiple places of deployment-charts CI causes various failures when checking charts/deployments
Change #1266216 had a related patch set uploaded (by Effie Mouzeli; author: Effie Mouzeli):
[operations/deployment-charts@master] Update fixtures and remove mw-parsoid
Change #1266216 merged by jenkins-bot:
[operations/deployment-charts@master] Update fixtures and remove mw-parsoid
Change #1265420 abandoned by Effie Mouzeli:
[operations/puppet@production] profile::service_proxy::envoy: remove mw-parsoid
Reason:
this was merged in I2c626f98f0b3627f8c690a7b0f5e94552decec73
