Last Updated on 2022-03-24 by Clay
今天我碰到一個情況，是伺服器中所安裝的 GitLab 無法正常啟動，且透過終端機使用 sudo systemctl start gitlab-runsvdir.service 會使畫面凍結（freeze）住。
無論使用：
sudo gitlab-ctl status
sudo gitlab-ctl stop
sudo gitlab-ctl start
都會出現：
fail: alertmanager: runsv not running
fail: crond: runsv not running
fail: gitaly: runsv not running
fail: gitlab-exporter: runsv not running
fail: gitlab-workhorse: runsv not running
fail: grafana: runsv not running
fail: logrotate: runsv not running
fail: nginx: runsv not running
fail: node-exporter: runsv not running
fail: postgres-exporter: runsv not running
fail: postgresql: runsv not running
fail: prometheus: runsv not running
fail: redis: runsv not running
fail: redis-exporter: runsv not running
fail: registry: runsv not running
fail: sidekiq: runsv not running
fail: unicorn: runsv not running
所以即便想要重新啟動 GitLab 也無法。
解決方法
或許有一種可能，是被 plymouth-quit-wait.service 這項行程給堵塞住了。這是參考自 https://forum.gitlab.com/t/plymouth-quit-wait-service-preventing-gitlab-runsvdir-service-possible/48442 這篇討論。發文者提到他花了兩小時在啟動他的 GitLab CE（我所遇到的也同樣是此版本），最後他鎖定了 plymounth-quit-wait.service 是問題的根源。
而我也是同樣的狀況。
我在使用以下指令後，會看到許多阻塞的行程：
sudo systemctl list-jobs
Output:
JOB UNIT TYPE STATE
98 gitlab-runsvdir.service start waiting
1 graphical.target start waiting
120 systemd-update-utmp-runlevel.service start waiting
182 ureadahead-stop.timer start waiting
104 plymouth-quit-wait.service start running
174 system-getty.slice start waiting
14 setvtrgb.service start waiting
2 multi-user.target start waiting
8 jobs listed.
而這些行程都在等待編號 104 的 plymouth-quit-wait.service。
接著我通過查詢網路上的說明，得知此行程僅只是在開機過程中讀取啟動的圖示、並應該在啟動結束卸載掉。
而我的伺服器自然是已經開機完成了，於是我毫無顧忌地使用以下指令結束掉該行程：
sudo systemctl stop plymouth-quit-wait.service
sudo systemctl start gitlab-runsvdir.service
sudo systemctl list-jobs
Output:
No jobs running.
確認沒有任何行程阻塞，接著嘗試啟動 gitlab：
sudo gitlab-ctl start
Output:
ok: run: alertmanager: (pid 28589) 31s
ok: run: crond: (pid 28570) 31s
ok: run: gitaly: (pid 28574) 31s
ok: run: gitlab-exporter: (pid 28571) 31s
ok: run: gitlab-workhorse: (pid 28586) 31s
ok: run: grafana: (pid 28588) 31s
ok: run: logrotate: (pid 28566) 31s
ok: run: nginx: (pid 28562) 31s
ok: run: node-exporter: (pid 28572) 31s
ok: run: postgres-exporter: (pid 28585) 31s
ok: run: postgresql: (pid 28565) 31s
ok: run: prometheus: (pid 28575) 31s
ok: run: redis: (pid 28569) 31s
ok: run: redis-exporter: (pid 28577) 31s
ok: run: registry: (pid 28578) 31s
ok: run: sidekiq: (pid 28584) 31s
ok: run: unicorn: (pid 28587) 31s這一次果然沒有任何問題了。
或許這是一個可以參考的解決步驟，很感謝當初留下紀錄的發文者。
References
- https://forum.gitlab.com/t/plymouth-quit-wait-service-preventing-gitlab-runsvdir-service-possible/48442
- https://askubuntu.com/questions/1119167/slow-boot-issue-due-to-plymouth-quit-wait-service-ubuntu-18-04