7.2 to 8.0 – Shutting down gitlab-git-http-server hangs
Created by: phanan
I'm using an installed-from-source Gitlab – Omnibus wasn't an option 3 years ago :( After upgrading from 7.2 to 8.0 today, there's this issue when I tried to (re)start Gitlab service:
$ service gitlab restart
Shutting down GitLab Unicorn
Shutting down GitLab Sidekiq
Shutting down gitlab-git-http-server
.......................................................(to infinity and beyond)^C
Session terminated, killing shell... ...killed.As you can see, the number of dots just keeps increasing to infinity. Now I'll kill it manually and start the service again:
$ service gitlab start
Removing stale gitlab-git-http-server pid. This is most likely caused by gitlab-git-http-server crashing the last time it ran.
Starting GitLab Unicorn
Starting gitlab-git-http-server
bundler: command not found: unicorn_rails
Install missing gem executables with `bundle install`
The Sidekiq job dispatcher is already running with pid 21214, not restarting
..............................Waited 30s for the processes to write their pids, something probably went wrong.
.........................Another try you say?
$ service gitlab start
Starting GitLab Unicorn
bundler: command not found: unicorn_rails
Install missing gem executables with `bundle install`
The Sidekiq job dispatcher is already running with pid 21214, not restarting
The gitlab-git-http-server is already running with pid 21214, not restarting
..............................Waited 30s for the processes to write their pids, something probably went wrong.And of course, nginx returns a beautiful 502 error.
Anyone can give me a clue what's going on here?