Package Details: seahub 7.0.4-1

Git Clone URL: (read-only)
Package Base: seahub
Description: The web frontend for seafile server
Upstream URL:
Licenses: Apache
Submitter: eolianoe
Maintainer: Joffrey
Last Packager: Joffrey
Votes: 6
Popularity: 0.000026
First Submitted: 2017-07-03 09:48
Last Updated: 2019-08-20 16:19

Pinned Comments

Joffrey commented on 2018-02-13 10:56


Prerequisites :

  • The seafile-server daemon must be stopped
  • You must use seafile user
  • You must be placed in /your/instance/seafile-server/

Make a seahub backup :

mv seahub seahub-old

Copy the new directories

cp -r /usr/share/seafile-server/{upgrade,seahub} ./

Run an upgrade :


Otherly you can use my script seafile-helper available on AUR.

Latest Comments

« First ‹ Previous 1 2 3 4 5 6 Next › Last »

kuzalj commented on 2018-01-29 01:23

@klemens Thank you so much! I wondered where those logs were. I found out it was erroring because I did not have the two factor tables in my sql database, which is odd because those got introduced back in 6.0.0 and i know i executed the upgrade script. I just re-ran the database upgrade and I finally am in Seahub for the first time in a while haha.

@Joffrey If i cleanly uninstall the pip installed Gunicorn and install with pacman, do you think i can use the original PKGBUILD to install/run seahub?

klemens commented on 2018-01-28 21:55

@kuzalj the seahub logs are stored in /tmp/seahub.log and /tmp/seahub_django_request.log (at least in my case).

kuzalj commented on 2018-01-28 21:30

@Joffrey that PKGBUILD did allow for Seahub to start. I get the login page. However when i try to log in i get a "Page unavailable Sorry, but the requested page is unavailable due to a server hiccup.

Our engineers have been notified, so check back later." error. But i think that is a seperate problem i will have to look into. Is there something i should do to fix the gunicorn issue or are there any logs you want? Thank you for the help thus far.

klemens commented on 2018-01-28 16:24

openjpeg2 is needed as a dependency, otherwise the pip-installed pillow cannot generate jpeg thumbnails (why do you use pip instead of just depending on existing packages in the official repos where possible?).

Joffrey commented on 2018-01-27 08:45

@Captain_Rage I had a similar problem. Can you try this:

chmod 755 /home/seafile
chown -R seafile:seafile /home/seafile

And run systemd service with seafile user, When you run the service as root you change the owner of some files.

Captain_Rage commented on 2018-01-27 01:02

The thing is, when I set User=seafile and Group=seafile in /etc/systemd/system/seafile-server@.service, systemctl status seafile-service@myinstance shows:

[root@Chimney chimney]# systemctl status seafile-server@longboat
* seafile-server@longboat.service - Next-generation open source cloud storage with advanced features on privacy 
   Loaded: loaded (/etc/systemd/system/seafile-server@.service; indirect; vendor preset: disabled)
   Active: active (running) since Fri 2018-01-26 19:06:10 CET; 2min 22s ago
  Process: 8136 ExecStop=/usr/bin/seafile-admin stop (code=exited, status=0/SUCCESS)
  Process: 8146 ExecStart=/usr/bin/seafile-admin start (code=exited, status=0/SUCCESS)
 Main PID: 8154 (seaf-server)
    Tasks: 14 (limit: 4915)
   CGroup: /system.slice/system-seafile\x2dserver.slice/seafile-server@longboat.service
           |-8151 seafile-controller -c /home/seafile/longboat/ccnet -d /home/seafile/longboat/seafile-data -F /
           |-8152 ccnet-server -F /home/seafile/longboat/conf -c /home/seafile/longboat/ccnet -f /home/seafile/l
           `-8154 seaf-server -F /home/seafile/longboat/conf -c /home/seafile/longboat/ccnet -d /home/s

But when User=seafile and Group=seafile are omitted (I suppose then it is run as root...?), systemctl will instead show:

CGroup: /system.slice/system-seafile\x2dserver.slice/seafile-server@longboat.service
           |-8040 seafile-controller -c /home/seafile/longboat/ccnet -d /home/seafile/longboat/seafile-data -F /
           |-8041 ccnet-server -F /home/seafile/longboat/conf -c /home/seafile/longboat/ccnet -f /home/seafile/l
           |-8043 seaf-server -F /home/seafile/longboat/conf -c /home/seafile/longboat/ccnet -d /home/seafile/lo
           |-8059 python2.7 /usr/lib/seafile/seafileenv/bin/gunicorn seahub.wsgi:application -c /home/seafile/lo
           |-8126 python2.7 /usr/lib/seafile/seafileenv/bin/gunicorn seahub.wsgi:application -c /home/seafile/lo
           |-8127 python2.7 /usr/lib/seafile/seafileenv/bin/gunicorn seahub.wsgi:application -c /home/seafile/lo
           `-8128 python2.7 /usr/lib/seafile/seafileenv/bin/gunicorn seahub.wsgi:application -c /home/seafile/lo

Now there are entries about seahub running. Why doesn't this work when the process is ran as the user seafile?

EDIT: Also, now I found that the manual actually states how to go about to create a systemd unit for seahub: It says to execute '${seafile_dir}/seafile-server-latest/ start' to start seahub. In the current state of how the packages look in the AUR, it still boggles my mind how seahub gets started and how it supposedly runs or should be run.

Joffrey commented on 2018-01-27 00:30

@kuzalj your pacman database is corrupt..

For seahub I think understand the problem.. If gunicorn is already installed on your system, when pip install gunicorn in the virtualenv created with --site-packages option, pip find the module gunicorn in virtualen.. So pip don't reinstall it, and you haven't "/usr/lib/seafile/seafileenv/bin/gunicorn" executable..

Can you try with this PKGBUILD, I use it for my instance for the moment. It would confirm my hypothesis.

kuzalj commented on 2018-01-26 19:25


I ran that and gunicorn was not installed as an Arch package. When I tried to install the package, it errored out citing file conflicts. This means that it must have gotten installed some other way (pip?). Either way, I have gunicorn and that doesn't seem to be the problem. What is weird in 6.2.5-1 it did start seahub fine, but now does not.

Joffrey commented on 2018-01-26 10:58

@Captain_Rage If I understand your problem correctly.. If you use seafile-server@.service provided by seafile-server pkg %i replace your INSTANCENAME and it is called with:

systemctl start seafile-server@INSTANCENAME.service

Very approximately systemd run seafile-admin, seafile-admin run gunicorn (wcgi) and you use a reverse proxy (Nginx or Apache) for converse with gunicorn.

Captain_Rage commented on 2018-01-25 14:54

How is seahub actually getting started? The systemd unit of seafile-server gives me no clue and I can't to figure in any way. They seem to start in tandem upon using systemctl start seafile-server.service. This is really confusig, since I'm trying to change an old configuration of seafile-server to run smoothly between upgrades (now that seahub has its own AUR package and FastCGI is no longer endorsed).

EDIT: I mean, systemctl start seafile-server.service starts seafile-server with '/usr/bin/seafile-admin start'. What about seahub?

EDIT2: Is seahub being started by the web server, once the user has added an entry for it there? E.g. nginx.