Project

General

Profile

Bug #741

Redmine: "Incomplete response received from application"

Added by Pierre-Louis Bonicoli 29 days ago. Updated 16 days ago.

Status:
Resolved
Priority:
Normal
Category:
Service :: Web
Start date:
2021-11-02
Due date:
% Done:

100%

Estimated time:
Patch Available:
Confirmed:
No
Branch:
Entity:
DuckCorp
Security:
Help Needed:

Description

I received this response twice today, both time at an issue creation.

<h2>Incomplete response received from application</h2>

It happens while updating an existing issue too. Using F5 allows to workaround this issue.

Associated revisions

Revision 5e72c249 (diff)
Added by Marc Dequènes 16 days ago

redmine: increase memory limit

fixes #741

History

#1

Updated by Pierre-Louis Bonicoli 29 days ago

  • Description updated (diff)
#2

Updated by Marc Dequènes 16 days ago

  • Status changed from New to In Progress
  • Assignee changed from Pierre-Louis Bonicoli to Marc Dequènes

I see these errors on the reverse proxy:

AH01102: error reading status line from remote server 10.0.7.2:80
AH00898: Error reading from remote server returned by /projects/dc-ducklings-volunteer-activities/issues.atom

In the container I see those:

[ 2021-11-16 05:13:05.9810 2586983/7f84aaffd700 age/Cor/Con/InternalUtils.cpp:112 ]: [Client 12-6864] Sending 502 response: application did not send a complete response
App 2831951 stdout: 
[ 2021-11-16 05:13:06.9367 2586983/7f8500a1c700 age/Cor/Con/InternalUtils.cpp:112 ]: [Client 2-6865] Sending 502 response: application did not send a complete response
[ 2021-11-16 05:13:07.2635 2586983/7f84eb7fe700 age/Cor/CoreMain.cpp:819 ]: Checking whether to disconnect long-running connections for process 2831931, application redmine_md

And those too:

[ 2021-11-16 01:06:24.8866 2586983/7f84e82b4700 age/Cor/Spa/SmartSpawner.h:740 ]: The application preloader seems to have crashed, restarting it and trying again...

I just found out the source of the problem:

[Tue Nov 16 06:17:13 2021] oom-kill:constraint=CONSTRAINT_MEMCG,nodemask=(null),cpuset=lxc.payload.redmine,mems_allowed=0-1,oom_memcg=/lxc.payload.redmine,task_memcg=/lxc.payload.redmine/system.slice/apache2.service,task=ruby,pid=2171217,uid=1065569
[Tue Nov 16 06:17:13 2021] Memory cgroup out of memory: Killed process 2171217 (ruby) total-vm:414356kB, anon-rss:129448kB, file-rss:8932kB, shmem-rss:0kB, UID:1065569 pgtables:444kB oom_score_adj:0
[Tue Nov 16 06:17:13 2021] oom_reaper: reaped process 2171217 (ruby), now anon-rss:0kB, file-rss:0kB, shmem-rss:0kB

I increased the memory limit for the container (the default stays at 512MB).

Current usage (after increase):

# lxc info redmine
…
  Memory usage:
    Memory (current): 620.11MiB
…

But it was limited to 512MB. Now we have a better understanding of how (too) much memory is being used by this service ;-P.

#3

Updated by Marc Dequènes 16 days ago

  • Status changed from In Progress to Resolved
  • % Done changed from 0 to 100

Also available in: Atom PDF