Worker Time Out

Dear Community,

Recently I installed seafile with seahub (6.1.1) on a Synology DS213j NAS station, that has a Debian Wheezy installation. Generally it works brilliantly and fast, that is, using the desktop client or iPhone app. Currently it is populated with project data of the last few episodes of our film group, including many image sequences.

Libraries / Files = 19 / 273500
Storage Used = 1.8 TB

The issue is as follows: The web-interface also works as expected, but only when DEBUG = True is appended to the conf/ Down/upload works, admin interface works, and much more.

When the setting DEBUG = False is set, or when it is omitted, the web-interface becomes completely irresponsive. I’m not using fastcgi (I would like to see it working without fastcgi before continuing with nginx).
The ./seafile-server/runtime/error.log shows that workers timeout:
2017-09-06 18:16:41 [27824] [INFO] Booting worker with pid: 27824
2017-09-06 18:18:42 [27490] [CRITICAL] WORKER TIMEOUT (pid: 27824)
2017-09-06 18:18:42 [27490] [CRITICAL] WORKER TIMEOUT (pid: 27824)
In here, 27490 is the pid of the main process. This pattern is repeated for a list of other gunicorn workers.

In the meantime I see that in /tmp/seahub_cache various .djcache-files are being touched. They do not seemingly change in size.

In /tmp/seahub_cache and in the location where seafile-admin is executed all files are (recusively) owned by seafile:seafile. The seafile-admin start is executed as the seafile user (however, in the past it was started as root, afterwards the /tmp/seahub_cache was flushed).

As I feel stuck with a couple of evenings and weekends spend on this problem, I turn my hopes to you.

My questions are the following:

  1. Do you know why the workers time-out? Has it something to do with one-time-caching?
  2. If so: Is it possible to extend the timeout duration to a couple of days? This may give time to create a one-time cache.
  3. If so: Are thumbnails being created during this process for all images and videos? When this is the case it can explain the timeout for such a slow NAS.
  4. Except for the django template debugging screen, what elements are disabled during the DEBUG = True? This may help me to find the cause…

Many thanks in advance,


  • Update: Minor grammer related stuff.

Have you configured Nginx + Memcached? Do you use MySQL as database?

It was not configured with Nginx + Memcached, and there was no mysql installed.

Amof let me to the right solution, however, it took me a while to realize that seahub.conf is located in ./seafile-server/runtime/seahub.conf instead of using ./conf/