(mod_fastcgi.c.1406) [ERROR]: spawning fcgi failed

[xenithorb]

2018-01-13 13:52:17: (log.c.164) server started
2018-01-13 13:52:17: (mod_fastcgi.c.984) bind failed for: unix:/var/run/lighttpd/php.socket-0 No such file or directory
2018-01-13 13:52:17: (mod_fastcgi.c.1406) [ERROR]: spawning fcgi failed.
2018-01-13 13:52:17: (server.c.1022) Configuration of plugins failed. Going down.

Getting this in /var/log/lighthttpd/error.log

docker logs shows s6 is restarting lighthttpd over and over again, but otherwise pi-hole setup was successful and DNSmasq is running and working. FTL is also running OK.

REPOSITORY                 TAG                 IMAGE ID            CREATED             SIZE
docker.io/diginc/pi-hole   latest              6b77d5329fd4        8 days ago          330.5 MB

[diginc]

Are there any volumes pointed to /var/? What's the container run configuration look like?

[jimsantora]

I'm seeing the same issue with that image, as well as the current debian_dev tag. Simply creating the /var/run/lighttpd directory inside the container and giving www-data r/w access fixed the problem.

Also tried building from your dockerfile, and the resulting image had the same problem. I didn't see any errors when lighttpd was installed, maybe something has changed with the deb. For now, adding a couple lines to the Dockerfile to create the directory and set perms works.

[tybritten]

+1 on this issue. should we put a PR in for the temp fix?

[diginc]

I've been out sick for a week, trying to reproduce with docker_run.sh now.

[diginc]

I was unable to reproduce. Can I see a docker run command example? Also what OS & Docker version are those of you seeing this problem using?

I ran this exactly: https://raw.githubusercontent.com/diginc/docker-pi-hole/master/docker_run.sh

$ docker inspect pihole | grep -i image
        "Image": "sha256:6b77d5329fd484527ec719e52e3f2c85cecea0a06a652feeed8779f458d8abd0",
            "Image": "diginc/pi-hole:latest",
                "image": "diginc/pi-hole:debian_amd64",

/var/run/lighttpd permissions:

$ docker exec pihole ls -lat /var/run/lighttpd
total 8
drwxr-xr-x 2 www-data www-data 4096 Jan 25 15:17 .
drwxr-xr-x 8 root     root     4096 Jan 25 15:17 ..
srwxr-xr-x 1 www-data www-data    0 Jan 25 15:17 php.socket-0

[tybritten]

docker inspect pihole | grep -i image
        "Image": "sha256:6b77d5329fd484527ec719e52e3f2c85cecea0a06a652feeed8779f458d8abd0",
            "Image": "diginc/pi-hole",
                "image": "diginc/pi-hole:debian_amd64",

and here:

docker exec pihole ls -lat /var/run/lighttpd
ls: cannot access /var/run/lighttpd: No such file or directory

Docker version 1.13.1, build fbadd78-unsupported

fedora-atomic-27:fedora/27/x86_64/atomic-host
                   Version: 27.47 (2018-01-04 01:48:32)
                    Commit: 397e907961adafaeff11b807ceade8da5783134072406fcdba627f1195e0db76
              GPGSignature: Valid signature by 860E19B0AFA800A1751881A6F55E7430F5282EE4

[tybritten]

whoops almost forgot the docker run:

/usr/bin/docker run --rm \
--name pihole \
--dns 127.0.0.1 --dns 8.8.8.8 \
-p 53:53/tcp -p 53:53/udp -p 80:80 \
-v "/var/opt/pihole/pihole/:/etc/pihole/" \
-v "/var/opt/pihole/dnsmasq.d/:/etc/dnsmasq.d/" \
-e ServerIP= ********** \
-e WEBPASSWORD= ******** \
-e VIRTUAL_HOST=********* \
diginc/pi-hole

[diginc]

what about docker exec pihole ls -lat /var/run/

#220 reports this issue happening on CentOS and you report Fedora. Since I'm unable to reproduce the problem on ubuntu I'm wondering if there's a connection between this bug and the Docker OS being RedHat based. @xenithorb what OS are you?

If the work around is simply a mkdir /var/run/lighttpd in install.sh I'm ok with that adding that...but I'm very curious what the cause is still.

[tybritten]

Interesting.
Here:

docker exec pihole ls -lat /var/run/
total 28
-rw-r--r-- 1 root root    4 Jan 25 17:38 lighttpd.pid
drwxr-xr-x 5 root root  220 Jan 25 17:38 .
-rw-r--r-- 1 root root    4 Jan 25 17:38 pihole-FTL.port
---------- 1 root root    0 Jan 25 17:38 crond.reboot
-rw-r--r-- 1 root root    3 Jan 25 17:38 pihole-FTL.pid
-rw-r--r-- 1 root root    4 Jan 25 17:38 crond.pid
-rw-r--r-- 1 root root    4 Jan 25 17:38 dnsmasq.pid
drwxr-xr-x 1 root root 4096 Jan 25 17:38 ..
drwxr-xr-x 7 root root  140 Jan 25 17:38 s6
drwxr-xr-x 2 root root   40 Jan 25 17:38 lock
drwxr-xr-x 3 root root 4096 Jan 25 17:38 secrets

so much for the host OS not being that important to container images..

[xenithorb]

CentOS 7.4 here @diginc

[diginc]

just-containers/s6-overlay#158 (comment) maybe the cause or related?

[tybritten]

Yeah looks like it- I think the best workaround for now is to add the VOLUME /run to the dockerfile or just make a note for people to add -v /run to the docker run command.

Thoughts?

[diginc]

I took away that renaming /init to something different would also be a valid work around since it would prevent the auto mounting /var.

[diginc]

debian_dev tag use /s6-init instead of /init now: https://hub.docker.com/r/diginc/pi-hole/tags/

[xenithorb]

I can report that both solutions work, mounting -v /run and also the renaming of /init

Thanks