Skip to content

ipv6nat: restart issue for other containers with php

I am currently testing the ipv6nat version.

I just noticed that when stopping a service based on php-fpm and restarting it, I have an error 502 Bad gateway In the logs I can see:

web_1    | 2020/04/13 22:18:16 [error] 7#7: *1 connect() failed (111: Connection refused) while connecting to upstream, client: 172.18.0.5, server: , request: "GET /apps/files/?dir=/&fileid=2 HTTP/1.1", upstream: "fastcgi://[fd00:dead:beef::9]:9000", host: "cloud.localhost"
web_1    | 2020/04/13 22:18:16 [warn] 7#7: *1 upstream server temporarily disabled while connecting to upstream, client: 172.18.0.5, server: , request: "GET /apps/files/?dir=/&fileid=2 HTTP/1.1", upstream: "fastcgi://[fd00:dead:beef::9]:9000", host: "cloud.localhost"
web_1    | 2020/04/13 22:18:16 [error] 7#7: *1 connect() failed (111: Connection refused) while connecting to upstream, client: 172.18.0.5, server: , request: "GET /apps/files/?dir=/&fileid=2 HTTP/1.1", upstream: "fastcgi://[fd00:dead:beef::8]:9000", host: "cloud.localhost"
web_1    | 2020/04/13 22:18:16 [warn] 7#7: *1 upstream server temporarily disabled while connecting to upstream, client: 172.18.0.5, server: , request: "GET /apps/files/?dir=/&fileid=2 HTTP/1.1", upstream: "fastcgi://[fd00:dead:beef::8]:9000", host: "cloud.localhost"
web_1    | 2020/04/13 22:18:16 [error] 7#7: *1 connect() failed (111: Connection refused) while connecting to upstream, client: 172.18.0.5, server: , request: "GET /apps/files/?dir=/&fileid=2 HTTP/1.1", upstream: "fastcgi://[fd00:dead:beef::6]:9000", host: "cloud.localhost"
web_1    | 2020/04/13 22:18:16 [warn] 7#7: *1 upstream server temporarily disabled while connecting to upstream, client: 172.18.0.5, server: , request: "GET /apps/files/?dir=/&fileid=2 HTTP/1.1", upstream: "fastcgi://[fd00:dead:beef::6]:9000", host: "cloud.localhost"
web_1    | 2020/04/13 22:18:16 [error] 7#7: *1 connect() failed (111: Connection refused) while connecting to upstream, client: 172.18.0.5, server: , request: "GET /apps/files/?dir=/&fileid=2 HTTP/1.1", upstream: "fastcgi://172.18.0.9:9000", host: "cloud.localhost"
web_1    | 2020/04/13 22:18:16 [warn] 7#7: *1 upstream server temporarily disabled while connecting to upstream, client: 172.18.0.5, server: , request: "GET /apps/files/?dir=/&fileid=2 HTTP/1.1", upstream: "fastcgi://172.18.0.9:9000", host: "cloud.localhost"
web_1    | 2020/04/13 22:18:16 [error] 7#7: *1 connect() failed (111: Connection refused) while connecting to upstream, client: 172.18.0.5, server: , request: "GET /apps/files/?dir=/&fileid=2 HTTP/1.1", upstream: "fastcgi://172.18.0.8:9000", host: "cloud.localhost"
web_1    | 2020/04/13 22:18:16 [warn] 7#7: *1 upstream server temporarily disabled while connecting to upstream, client: 172.18.0.5, server: , request: "GET /apps/files/?dir=/&fileid=2 HTTP/1.1", upstream: "fastcgi://172.18.0.8:9000", host: "cloud.localhost"
web_1    | 2020/04/13 22:18:16 [error] 7#7: *1 connect() failed (111: Connection refused) while connecting to upstream, client: 172.18.0.5, server: , request: "GET /apps/files/?dir=/&fileid=2 HTTP/1.1", upstream: "fastcgi://172.18.0.6:9000", host: "cloud.localhost"
web_1    | 2020/04/13 22:18:16 [warn] 7#7: *1 upstream server temporarily disabled while connecting to upstream, client: 172.18.0.5, server: , request: "GET /apps/files/?dir=/&fileid=2 HTTP/1.1", upstream: "fastcgi://172.18.0.6:9000", host: "cloud.localhost"
web_1    | 172.18.0.5 - - [13/Apr/2020:22:18:16 +0000] "GET /apps/files/?dir=/&fileid=2 HTTP/1.1" 502 157 "-" "Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:75.0) Gecko/20100101 Firefox/75.0" "172.18.0.1"

If I remove the docker network web and recreate it without the option --ipv6 --subnet=fd00:dead:beef::/48, I cannot reproduce the issue.

I will continue my test, because the behaviour is strange:

  • When I start all services the first time, it is working fine
  • If I stop one service (e.g.: cloud) with docker-compose down, and restart it, then I have the error 502
  • If I stop all services including traefik (with docker-compose down) and restart them all it seems to work
  • If I stop cloud with docker-compose stop and restart it, the I have also the error 502
  • If I stop cloud and traefik with docker-compose down and restart them, I still have the error 502
  • If I stop the service and traefik and restart them, the 502 error is still there

the IP of the app container (the one containing php-fpm) is: 172.24.0.4 while the error message shows that the web container is trying to connect to it on 172.18.0.6.

It seems to happen for all php application behind an nginx reverse proxy. I cannot reproduce the issue with application containing the web server in the app container (like wallabag or kanboard).

Did you face a similar issue ? Any idea of any investigation that could enable to deep dive this issue ?

Edited by doc75