lighttpd breaks after large system update (kind of solved)

luk32 · 02-24-2010, 11:14 AM

I am not fully sure if this belongs more to the networking section or the software. O hope it is related to here enough.

My problem is as the topic says.

My port 80 seems to be already taken. lighttpd fails to start reporting that it cannot bind to port cause its already taken.
The error message itself:

Code:

"Starting web server: lighttpd2010-02-21 07:01:02: (network.c.345) can't bind to port: :: 80 Address already in use
 failed!"

I am not able to crack this by myself. Ihve tried investigating this by

Code:

lsof -i :80
netstat -ape | grep -v unix

First doesn't show anything second anything abt port 80.

Now hints I can think of. Maybe you have more experience and ideas what I can do and where should I look for the culprit.

I am using debian (unstable) and lighttpd. I even tried rebooting; it says that the port is taken already at the boot sequence (well it starts pretty much at end anyway but the system should be clean).

It seems that it started failing after a major distro update, so maybe there are some conflicts based on ipv6/v4 or something, though I have no idea what it could be. Well my hit for that is the part of error msg "bind to port: :: 80 ". This "::" looks like ipv6 naming convention. Maybe its lighttpd internal.

If you have any ideas what it is or how can I check please help

.

Cheers.

nimnull22 · 02-24-2010, 11:17 AM

what "netstat -ntulp" from root console gives you?

luk32 · 02-24-2010, 11:40 AM

Well pretty much the same as "netstat -ape | grep -v unix". Nothing related to port 80.

Though it seems that the problem is related to lighttpd specifically. I have started tampering with config file and found lines:

Code:

######### Options that are good to be but not neccesary to be changed #######
## Use ipv6 only if available. (disabled for while, check #560837)
include_shell "/usr/share/lighttpd/use-ipv6.pl"

I've commented it out and it worked. So it is at least a partial solution.
Dunno what it does but it seems that after upgrade of lighttpd this line breaks validity of the config file and causes lighttpd to bind to port 80 twice.
That is why the port seems to be miraculously taken when system tools report nothing suspicious.

I also checked that when I remove the line it seems that lighttpd does not respond on ipv6 calls.

Code:

luk32debian:/etc/lighttpd# wget -6 ip6-localhost
--2010-02-21 12:42:46--  http://ip6-localhost/
Resolving ip6-localhost... ::1
Connecting to ip6-localhost|::1|:80... failed: Connection refused.

Never the less I accept my self-criticism for taking your time, obviously I could have done more by myself. Also the topic should be probably rebranded and moved to more appropriate section as the issue is lighttpd/config specific. Maybe the newer version is broken by the mentioned line or maybe combined with ipv6 turned on breaks it.

Thanks for reading into my issue anyway

.

Cheers.