On Tue, Mar 03, 2015 at 12:12:44PM +0100, Michael Gebetsroither wrote:
Package: nginx-full
Version: 1.6.2-5~bpo70+1
Severity: normal
Dear Maintainer,
After upgrading to nginx-full from backports `service nginx
reload/restart` does not work anymore.
17:20:17 [alert] 502#0: *6580993 open socket #13 left in connection 65
17:20:17 [alert] 502#0: *3806892 open socket #25 left in connection 186
17:20:17 [alert] 502#0: *2082012 open socket #24 left in connection 202
17:20:17 [alert] 502#0: aborting
The non-working reload is unfortune but not that bad. The really bad one
is restart not restarting nginx but only "stopping" it but not starting
it again.
It would be nice to have restart force fully starting nginx if stopping
it has failed, wouldn't it?
Currently we just use nginx stop; killall -9 nginx; nginx start to have
nginx reliable started again.
ps.: this is on a quite low traffic site with SPDY and SSL activated.
pps.: the system information below is fixed by hand and bug is sent from
my dev machine, but the information should be correct nevertheless.
greets,
Michael
Hello Michael,
I wasn't able to reproduce this behaviour. I tried reloading and
restarting nginx, while at the same time, I kept an http and a spdy
connection open.
Everything worked as expected, the connection was terminated after 5
secs (the recent graceful stop change) and nginx was restarted
successfully.
Are you sure you are using the latest /etc/init.d/nginx and
/etc/default/nginx? Make sure the STOP_SCHEDULE variable is set
correctly.
If everything is ok, could you provide some steps to reproduce it?
thanks for your bug report,
chris
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
