I'm looking at it now: I have a 502 like yesterday but it's not a
problem of disk space this time.

I found this error:
2016/04/23 11:22:36 [error] 21885#0: *206 connect() to
unix:/var/opt/gitlab/gitlab-workhorse/socket failed (111: Connection
refused) while connecting to upstream, client: 151.80.31.175, server:
git.devuan.org, request: "GET
/d-i/base-installer/blob/5adff8e4e4fd8acda27ee973ec31ce3fe7965787/kernel
/m68k.sh
HTTP/1.1", upstream:
"http://unix:/var/opt/gitlab/gitlab-workhorse/socket:/d-i/base-installer
/blob/5adff8e4e4fd8acda27ee973ec31ce3fe7965787/kernel/m68k.sh",
host: "git.devuan.org"

The socket file of this component exists:
ls -al /var/opt/gitlab/gitlab-workhorse/socket
srwxrwxrwx 1 git git 0 Apr 22 18:01
/var/opt/gitlab/gitlab-workhorse/socket

even if the daemon start datetime is different:
ps aux | grep workh
root      2785  0.0  0.0   3964    84 ?        Ss   Mar13   0:00 runsv
gitlab-workhorse
root      2788  0.0  0.0   4108   188 ?        S    Mar13   0:38
svlogd -tt /var/log/gitlab/gitlab-workhorse
git      21866  0.0  0.1  58548  3588 ?        Ssl  11:19   0:00
/opt/gitlab/embedded/bin/gitlab-workhorse -listenNetwork unix
- -listenUmask 0 -listenAddr /var/opt/gitlab/gitlab-workhorse/socket
- -authBackend http://localhost:8080 -authSocket
/var/opt/gitlab/gitlab-rails/sockets/gitlab.socket -documentRoot
/opt/gitlab/embedded/service/gitlab-rails/public -pprofListenAddr

and the last entry in the logfile id of yesterday afternoon:

2016-04-22_16:24:43.32956 git.devuan.org @ - - [2016-04-22
18:24:43.199317037 +0200 CEST] "GET
/devuan-infrastructure/jenkins-debian-glue/blob/8279e905800a8a84e6ea3333
bcc9308e19f8a883/debian/jenkins-debian-glue.install
HTTP/1.1" 200 18643 "" "Mozilla/5.0 (compatible; MJ12bot/v1.4.5;
http://www.majestic12.co.uk/bot.php?+)" 0.130178

I also see a lot of logging service with an old start date:
root      2784  0.0  0.0   4116   120 ?        Ss   Mar13   0:34
runsvdir -P /opt/gitlab/service log: lab/gitlab-workhorse: file does
not exist svlogd: pausing: unable to rename current:
/var/log/gitlab/unicorn: file does not exist svlogd: pausing: unable
to rename current: /var/log/gitlab/redis: file does not exist svlogd:
pausing: unable to rename current: /var/log/gitlab/postgresql: file
does not exist svlogd: pausing: unable to rename current:
/var/log/gitlab/sidekiq: file does not exist
root      2787  0.0  0.0   3964   148 ?        Ss   Mar13   0:00 runsv
unicorn
root      2789  0.0  0.0   4108   252 ?        S    Mar13   0:00
svlogd -tt /var/log/gitlab/unicorn

Yesterday I moved /var/log on the data disk but I didn't restart any
service: I simply saw the website up again, so I assumed that someone
else did the job.
My knowledge of last versions of GitLab is really limited, sorry but I
don't know how to solve the issue: reboot is still the last resort and
I can do it, but if someone with a deeper knowledge wants to take a
look, I think it's better.

I'm even more convinced that now we need a better setup with two VM in
two different zones in a "Nextime config", to be resilient in case of
this kind of issues.

Alberto

On 23/04/16 10:08, Jaromil wrote:
>
> hi
>
> anyone has clues why this happens again? gitlab stopped working. I
> was just running a search BTW. perhaps searches should be limited
> to registered users with some privilege?
>
> I feel insecure on what is happening on the gitlab side of things
> if there is an intervention please first notice this list ok?
>
> ciao _______________________________________________ devuan-dev
> internal mailing list devuan-dev@???
> https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/devuan-dev
>