Page MenuHomePhabricator

Parsoid crashing with cancel after 0 retries! message
Closed, DeclinedPublic

Description

While lurking through the RESTBase logs I've found some Parsoid crashes that weren't happening before:

Request: curl -H 'Cache-Control: no-cache' https://ru.wikipedia.org/api/rest_v1/page/html/%D0%A3%D1%87%D0%B0%D1%81%D1%82%D0%BD%D0%B8%D0%BA%3A%D0%9B%D0%B5_%D0%9B%D0%BE%D0%B9%2F%D0%9A%D0%BE%D1%82%D0%B5%D0%B3%D0%BE%D1%80%D0%B8%D1%87%D0%B5%D1%81%D0%BA%D0%B82?redirect=false

Error:

cancel after 0 retries!
ProcessTerminatedError: cancel after 0 retries!
    at Farm.<anonymous> (/srv/deployment/parsoid/deploy-cache/revs/1367057edce8c4813aa60b5715e5219a32c82611/node_modules/worker-farm/lib/farm.js:87:25)
    at Array.forEach (native)
    at Farm.<anonymous> (/srv/deployment/parsoid/deploy-cache/revs/1367057edce8c4813aa60b5715e5219a32c82611/node_modules/worker-farm/lib/farm.js:81:36)
    at ontimeout (timers.js:386:14)
    at tryOnTimeout (timers.js:250:5)
    at Timer.listOnTimeout (timers.js:214:5)

It's not a high priority, there's only about 600 messages like that per day

Related Objects

Event Timeline

mobrovac renamed this task from Parsoid crasher with cancel after 0 retries! message to Parsoid crashing with cancel after 0 retries! message.Feb 13 2018, 7:22 PM
mobrovac updated the task description. (Show Details)
Arlolra subscribed.

cancel after 0 retries!

This a generic message after,
https://github.com/wikimedia/parsoid/commit/481da3d5d83acdf5c5a1193d7c5aa720a7f24f20

and is unfortunately not very helpful in itself in diagnosing a problem.

It can be from uncaught exceptions, but also from Parsoid workers giving up from exceeding limits and various other things.

This particular instance is from,
https://logstash.wikimedia.org/app/kibana#/doc/logstash-*/logstash-2018.02.13/parsoid?id=AWGQlZQVqJPajYV1QDV9&_g=(refreshInterval:(display:Off,pause:!f,value:0),time:(from:now-24h,mode:quick,to:now))

<code>
  PHP fatal error: <br/>
  request has exceeded memory limit</code>

It is however helpful to monitor the rate occurrence. Increasing by an order of magnitude might indicate a pressing issue.