Closed tlc closed 9 years ago
If I kill the stuck node process, the baseimage stuff restarts it and it works. Perhaps the app is being started before some other part of the container is ready?
Hmm... 14.04 is my environment as well, and I haven't had any issues like this. I have seen issues where my devices were caching DNS and causing headaches. Is it possible that was your problem? You might try turning your wifi off and on again on your iPhone after airsonos starts up and see if that helps.
The not-seen/seen experience was the same on my iPhone and my Macbook Pro, but I'll try turning it off and on again tonight.
BTW, my wife's iPhone saw the duplicate entries I've seen mentioned elsewhere, but my devices did not.
The ps outputs from two different container runs show that the runsv's all run before any of their children start. But the child processes are started in different orders in the two runs.
# ps -ef
UID PID PPID C STIME TTY TIME CMD
root 1 0 0 15:37 ? 00:00:00 /usr/bin/python2 -u /sbin/my_init
root 13 1 0 15:37 ? 00:00:00 /usr/bin/runsvdir -P /etc/service log: ge pipeline;?Error opening file
root 14 13 0 15:37 ? 00:00:00 runsv cron
root 15 13 0 15:37 ? 00:00:00 runsv dbus
root 16 13 0 15:37 ? 00:00:00 runsv airsonos
root 17 13 0 15:37 ? 00:00:00 runsv syslog-ng
root 18 13 0 15:37 ? 00:00:00 runsv avahi
root 19 13 0 15:37 ? 00:00:00 runsv sshd
avahi 20 18 0 15:37 ? 00:00:00 avahi-daemon: running [mastershake.local]
103 22 15 0 15:37 ? 00:00:00 dbus-daemon --system --nofork
root 24 14 0 15:37 ? 00:00:00 /usr/sbin/cron -f
nobody 25 16 1 15:37 ? 00:00:00 node /usr/bin/airsonos
# ps -ef
UID PID PPID C STIME TTY TIME CMD
root 1 0 0 15:51 ? 00:00:00 /usr/bin/python2 -u /sbin/my_init
root 12 1 0 15:51 ? 00:00:00 /usr/bin/runsvdir -P /etc/service log: ge pipeline;?Error opening file
root 13 12 0 15:51 ? 00:00:00 runsv cron
root 14 12 0 15:51 ? 00:00:00 runsv dbus
root 15 12 0 15:51 ? 00:00:00 runsv airsonos
root 16 12 0 15:51 ? 00:00:00 runsv syslog-ng
root 17 12 0 15:51 ? 00:00:00 runsv avahi
root 18 12 0 15:51 ? 00:00:00 runsv sshd
root 19 13 0 15:51 ? 00:00:00 /usr/sbin/cron -f
103 20 14 0 15:51 ? 00:00:00 dbus-daemon --system --nofork
nobody 22 15 1 15:51 ? 00:00:00 node /usr/bin/airsonos
avahi 24 17 0 15:51 ? 00:00:00 avahi-daemon: running [mastershake.local]
I've created a docker build tagged as 'testing' against the runit_deps ( #2 ) branch. Can you give that a shot, and report back if that works?
It didn't work. But a 'sleep 5' in airsonos/run did.
Hmm. Likely mdns registers as "started" before it actually is. I'm not above a little hack that works... I'll push a commit with a sleep later this weekend.
Thanks for the help on this.
On Friday, May 22, 2015, tlc notifications@github.com wrote:
It didn't work. But a 'sleep 5' in airsonos/run did.
— Reply to this email directly or view it on GitHub https://github.com/justintime/docker-airsonos/issues/1#issuecomment-104833453 .
I added the sleep 5, committed, and merged. Docker builds are updating now. Thanks again for helping identify this. If for some reason this doesn't fully resolve the problem, let me know and I'll reopen it.
Running on Ubuntu 14.04, the airsonos process starts.
ps shows the node process up:
But they don't show up on my OSX or iOS device lists.
If I exec into the container and run airsonos there, they do show up in my device lists.
And ps shows node running twice. So the first one is hung?