After installing Columbus 2.9.1 on RHEL 6.10 the Celery service is not starting correctly.
In this case, it does not display any message on /etc/init.d/columbus start however when you do /etc/init.d/columbus status it will throw a warning.
The sympton is that Columbus login page won't let you login with a message like: "not able to login, contact your administrator"
The most likely root cause is that columbus_webapp database is missing some relations or that /var/run/celery directory has incorrect permissions. Expected permissions should look like:
$ ls -al /var/run/celery
[root@RHEL6 ~]# ls -al /var/run/celery/ total 32
drwxr-xr-x. 2 columbus root 4096 Oct 5 19:25 .
drwxr-xr-x. 33 root root 4096 Oct 1 10:17 ..
-rw-r--r--. 1 columbus columbus 89 Oct 5 19:22 celerybeat-schedule.bak
-rw-r--r--. 1 columbus columbus 2874 Oct 5 19:25 celerybeat-schedule.dat
-rw-r--r--. 1 columbus columbus 89 Oct 5 19:25 celerybeat-schedule.dir
-rw-r--r--. 1 columbus columbus 5 Oct 1 10:17 image-service.pid
-rw-r--r--. 1 columbus columbus 5 Oct 1 10:17 long-tasks.pid
-rw-r--r--. 1 columbus columbus 5 Oct 1 10:17 periodic-tasks.pid
-rw-r--r--. 1 columbus columbus 5 Oct 1 10:17 periodic-tasks.pid
Solution consists of removing the db and reinstalling the columbus-webapp rpm:
$ su - columbus
$ sudo /etc/init.d/columbus stop
$ dropdb columbus_webapp
$ sudo rpm -e --nodeps Columbus-webapp
$ sudo yum install Columbus-webapp
$ sudo /etc/init.d/columbus start
After this Celery should be Ok and login process should work as expected
.
Regards,
Arturo