sillelien / base-alpine Goto Github PK

Context

• Just switched to running our Go apps as a static binary in a container based off sillelien/base-alpine.
• Apps are running in containers on Docker Cloud (Ubuntu 14.04 nodes).

Problem

• When one of our services starts, it attempts to connect to a RDS instance.
• This was timing out.

Solution

• Add a 20 second wait before trying to connect.

Logs

...

DNS : Starting monitoring loop
dial tcp xx.xx.xx.x:xxxx: getsockopt: connection timed out
/data-api exited 1

I just updated my Dockerfile to sillelienbase-alpine:latest-92.
My services fail to resolve any hostname now.
I confirmed that something is indeed broken with dnsmasq by opening a console to the running container.

Connectivity OK:

# ping 8.8.8.8
PING 8.8.8.8 (8.8.8.8): 56 data bytes
64 bytes from 8.8.8.8: seq=0 ttl=51 time=26.564 ms
64 bytes from 8.8.8.8: seq=1 ttl=51 time=26.606 ms
64 bytes from 8.8.8.8: seq=2 ttl=51 time=26.629 ms

dnsmasq is not resolving any hostname:

# nslookup www.google.com 127.0.0.1
Server:    127.0.0.1
Address 1: 127.0.0.1 localhost

nslookup: can't resolve 'www.google.com': Try again

The config files look good though. Strange.

# cat /etc/dnsmasq-resolv.conf
nameserver 8.8.8.8
nameserver 8.8.4.4
search 624cd394-a36f-417a-99ea-d8c7b770a4ff.local.tutum.io

# cat /etc/resolv.conf
nameserver 127.0.0.1

dnsmasq is indeed running:

# ps aux
PID   USER     TIME   COMMAND
    1 root       0:00 s6-svscan -t0 /var/run/s6/services
   21 root       0:00 s6-supervise s6-fdholderd
  139 root       0:00 s6-supervise dns-hack
  140 root       0:00 s6-supervise dns
  141 root       0:00 s6-supervise haproxy
  142 root       0:00 s6-supervise syslog
  144 root       0:00 syslogd -n -D -O /dev/stdout
  147 root       0:00 sh ./run
  151 root       0:01 dnsmasq --expand-hosts --resolv-file=/etc/dnsmasq-resolv.conf --addn-hosts=/etc/hosts.links --no-daemon
  152 root       0:00 logger
  153 root       0:00 sh ./run
  520 root       0:00 /bin/sh
  656 root       0:00 /usr/local/sbin/haproxy -f /etc/haproxy/haproxy.conf
  668 root       0:00 sleep 30
  669 root       0:00 ps aux

Is there something that can be added to the run script?

Hi!

In the readme > keep it small section, there's a sample build script that recommends running

apk upgrade
apk update

I'm new to apk, but that looks backwards to me. Wouldn't you want to update the repositories before upgrading packages?

I am trying to get a domain-like resolution of DNS names, but i am probably configuring something wrong.

what i want is that for each container in the Tutum stack, "" and "." will be resolved to the correct IPs. for now, i seem to manage only getting /etc/hosts to have an entry for the container name, if anything.

it seems like the DNS hack is depending on some env variables (TUTUM_REST_HOST, etc.) that are not documented in the README. can you document what i need to set each to be in the Tutum stack stackfile?
and is what i am looking for even possible? i am building an email solution in which each service (postfix, postgresql, dovecot) run in a container, and these should talk to one another.

Hey is round robind DNS supposed to be working? Or is that a limitation of this work around