On Mon, Mar 09, 2015 at 10:58:52PM +0530, shirish शिरीष wrote:
at bottom :-
On 3/9/15, Ivan Baldo <iba...@adinet.com.uy> wrote:
Ok, decided to check current Debian Policy Manual and there isn't a
specific mention of what to do when starting a service after dpkg's
configuration phase and that service fails.
So I guess is up to maintainers of this package to decide.
For what is worth, I vote for just continuing anyway and let the
sysadmin fix it later, since it surely will see that the webserver isn't
starting...
Thanks for your consideration!
El 09/03/15 a las 10:14, Ivan Baldo escibió:
Hello Shirish!
Good news, thats it, you don't have IPv6 at all configured, not
even for the loopback interface, which is a pretty non standard setup
nowadays so be careful, other services could act badly.
I don't know what the Policy document says about this situation,
but shouldn't the install succeed even if Nginx couldn't be started?
For example, suppose that I want to remove all IPv6 support from a
server, I think the package should install correctly anyway and let me
fix the configuration later myself to avoid listening on IPv6 and
listen only in IPv4.
Or what happens if I have something else listening on port 80 (a
reverse proxy for example) and I want to install Nginx and then change
it to listen on port 81 instead?
This is not a bug on _installation_ but on _startup_, there's a
difference.
Of course this is my own personal opinion and haven't read the
Policy document in years...
Thanks all, have a great day everybody!
Hi all,
Can everybody (i.e. Ivan, Norbert and Kartik) can you all reproduce
the issue if IPv6 is not set does the installation and configuration
fail in that condition. It would be nice if people see if they run
into the same issue, then we found the root cause.
@Kartik, please see if the issue happens if IPv6 is not configured. If
it is please re-word the bug title, rework the severity if you think
it should not be an RC bug or/and patch the package.
This is definitely the problem, nginx cannot bind an ipv6 socket so it
fails to start.
The package stays in unconfigured state when nginx could not be started,
this is standard behaviour: dh_installinit by default adds a stanga in
the postinst script that runs `invoke-rc.d <service> start || exit $?`.
So when the service fails the script exits with a non-zero exit status.
Please let me know what do you want to do so I know what action I need
to take my end -
a. Wait for you to package a patched version.
b. Think of how I can get IPv6 only on the loopback interface 'lo' .
This is how I had done it
https://wiki.debian.org/DebianIPv6#How_to_turn_off_IPv6
basically
~$ cat /etc/default/grub | grep ipv6
GRUB_CMDLINE_LINUX_DEFAULT="ipv6.disable=1 init=/lib/systemd/systemd"
You can delete the `listen` ipv6 line from your nginx config (probably
/etc/nginx/sites-enabled/default) and run an `apt-get install` to
reconfigure the package. That should fix the issue for you.
--
Regards,
Shirish Agarwal शिरीष अग्रवाल
My quotes in this email licensed under CC 3.0
http://creativecommons.org/licenses/by-nc/3.0/
http://flossexperiences.wordpress.com
EB80 462B 08E1 A0DE A73A 2C2F 9F3D C7A4 E1C4 D2D8
--
To UNSUBSCRIBE, email to debian-bugs-rc-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
