10

This is the third time I install gitlab on server and never had issues before.

I decide to reinstall gitlab-ce ( 12.10.0 ) after dealing with issues on 12.9.0.

Before installing the new version, I had ensured that all gitlab files where removed and reboot the server.

Step1: Uninstall gitlab 12.9.0

# gitlab-ctl stop
# gitlab-ctl uninstall
# apt-get remove --purge gitlab-ce
# find / -iname "gitlab" then try to remove all outputed files
# init 6

Step 2: Fresh install gitlab 12.10.0

# apt-get update
# apt-get install gitlab-ce=12.10.0-ce.0

Then I updated my /etc/gitlab/gitlab.rb file according to my server config.

Finally, I executed:

gitlab-ctl reconfigure

The processes was blocked on :

ruby_block[wait for redis service socket] action run

Any solutions ?

madago
  • 129
  • 1
  • 13
  • Did you solve this in he meantime? Have the exact same problem but no solution anywhere. [This](https://forum.gitlab.com/t/gitlab-ce-rhel7-reconfigure-get-stuck-on-first-reconfigure/36422/1) has more additional information but no solution either. – Pratched Nov 03 '20 at 12:05
  • Yes but I didn't remember exactly what I've done.. Have a look at [Here](https://serverfault.com/questions/711566/redis-server-does-not-create-socket-file) – madago Nov 20 '20 at 20:28
  • Had the same Problem. This resolved it for me: https://docs.gitlab.com/omnibus/common_installation_problems/#reconfigure-freezes-at-ruby_blocksupervise_redis_sleep-action-run – Manuel Richarz Dec 01 '20 at 02:25

2 Answers2

36

The link provided by Manuel Richarz indicated there might be different lines upon which the installation may hang, one of them being the ruby_block[wait for redis service socket] action run line your question is about.

Their proposed solution is to open a new terminal and run:

systemctl restart gitlab-runsvdir.service

This worked for me for the ruby_block[wait for logrotate service socket] action run line.

Maximilian brutus III
  • 643
  • 9
  • 21
  • 1
    Saved me almost two hours of waiting on `ruby_block[wait for logrotate service socket] action run` – Andrew Mititi Nov 18 '21 at 12:37
  • Don't know if it was relevant but I deleted /var/opt/gitlab/redis/dump.rdb before restarting gitlab-runsvdir, based on this comment : https://gitlab.com/gitlab-org/omnibus-gitlab/-/issues/293#note_92244520 – JJP Feb 28 '22 at 11:04
10

7 years later, and the systemctl restart gitlab-runsvdir.service command in a secondary terminal did the trick quite well. Directly after executing this command, install no longer hangs and continued and finished all well.

Robin van Rooijen
  • 101
  • 1
  • 2