[
https://issues.apache.org/jira/browse/TIKA-3441?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17360016#comment-17360016
]
Cristian Zamfir commented on TIKA-3441:
---------------------------------------
A side question - if the parent process only monitors the child, I presume it's
safe to assume that it does not need a lot of memory. Can I assume I only need
200MB for the parent?
A pause after restart on bind issues looks like a good thing to have in any
case. But the child restarts successfully e.g., on OOM without hitting this
issues.
For a docker setup I can solve it with a liveness check on the /status api
endpoint, but it would be nice to also know more about the underlying issue
causing this.
> tika server stuck in loop trying to bind
> ----------------------------------------
>
> Key: TIKA-3441
> URL: https://issues.apache.org/jira/browse/TIKA-3441
> Project: Tika
> Issue Type: Bug
> Components: docker, server
> Affects Versions: 1.25
> Environment: {{PID TTY STAT TIME COMMAND}}
> {{ 1 ? Ssl 2:54 java -jar /tika-server-1.25.jar -h 0.0.0.0 -spawnChild
> TIKA_CHILD_JVM_OPTS=-JXmx3g -JXX:+ExitOnOutOfMemoryError -status}}
> {{ 127 ? Z 1:15 [tesseract] <defunct>}}
> {{ 131 ? Z 1:31 [tesseract] <defunct>}}
> {{ 139 ? Z 0:16 [tesseract] <defunct>}}
> {{ 219 ? Z 40:20 [tesseract] <defunct>}}
> {{ 324 ? Z 2:32 [tesseract] <defunct>}}
> {{ 342 ? Z 0:09 [tesseract] <defunct>}}
> {{ 343 ? Z 0:09 [tesseract] <defunct>}}
> {{ 380 ? Z 0:00 [tesseract] <defunct>}}
> {{ 430 ? Z 9:18 [tesseract] <defunct>}}
> {{ 435 ? Z 4:52 [tesseract] <defunct>}}
> {{ 446 ? Z 18:34 [tesseract] <defunct>}}
> {{ 453 ? Z 22:33 [tesseract] <defunct>}}
> {{ 461 ? Z 0:25 [tesseract] <defunct>}}
> {{ 517 ? Z 4:35 [tesseract] <defunct>}}
> {{ 526 ? Z 0:04 [tesseract] <defunct>}}
> {{ 536 ? S 0:00 36:39}}
> {{ 881708 pts/0 Ss 0:00 bash}}
> {{ 883782 ? Sl 0:00 java -XX:+ExitOnOutOfMemoryError -Djava.awt.headless=true
> -cp /tika-server-1.25.jar org.apache.tika.server.TikaServerCli -h 0.0.0.0
> TIKA_CHILD_JVM_OPTS=-JXmx3g -status --id
> ab854166-0a4b-4ecc-93a5-e4504d11e4e8}}
> {{ 883801 pts/0 R+ 0:00 ps ax}}
> Reporter: Cristian Zamfir
> Priority: Major
> Attachments: bind_issue.txt
>
>
> Tika seems to be stuck in a loop trying to bind. Please see the attached logs.
> It is also using a lot of CPU while doing that, so in an environment with
> horizontal scaling it causes other containers to spin up.
> This seems to happen from time to time to one of the multiple containers
> running, normally everything runs smoothly.
> Not sure why it fails to bind in the first place (it's running in docker).
> I suspect that it may be related to the fact that it is trying to bind very
> quickly.
> Ideas on how to debug this issue further are welcome.
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
