munin-users

[munin-users] Cannot assign requested address

[Adrian G. <ad...@ad...>]

Hello,

I got on some of my servers "Cannot assign requested address" after booting the box.
After manual restart of munin-node everything works.

Logs:

2010/02/25-09:13:14 MyPackage (type Net::Server::Fork) starting! pid(1592)                   
Binding to TCP port 4949 on host 10.40.0.26                                                  
2010/02/25-09:13:14 Can't connect to TCP port 4949 on 10.40.0.26 [Cannot assign requested address]
  at line 88 in file /usr/share/perl5/Net/Server/Proto/TCP.pm                                     
2010/02/25-09:13:14 Server closing!                                                               
shutdown() on closed socket GEN0 at /usr/lib/perl/5.8/IO/Socket.pm line 242.                      
Process Backgrounded                                                                              
2010/02/25-10:37:16 MyPackage (type Net::Server::Fork) starting! pid(11587)                       
Binding to TCP port 4949 on host 10.40.0.26   

Etch & Lenny (Debian 4 & 5) are installed on the production nodes, on Debian 4 boxes there is: ii  munin-node               1.2.5-1, Debian 5 got 1.2.6-10~lenny1. Both installed from debian repositories.

I found few links related to this issue, but there was no solution for me left.
https://bugs.launchpad.net/ubuntu/+source/munin/+bug/442575
http://osdir.com/ml/db.rddtool.munin.general/2007-09/msg00019.html
https://lists.ubuntu.com/archives/ubuntu-server-bugs/2008-August/004700.html

I checked the init.d order, but the network is already up while starting munin. 

My setup is munin from debian repositories driven over VPN network. I'm able to ping my VPN gateway while munin is telling me "Cannot assign requested address".

After rebooting the maschine again, I got munin status: * Munin-Node status unknown. and in the logfile there are errors mentioned above.



Regards,

Adrian

Re: [munin-users] Cannot assign requested address

[Lupe C. <lu...@lu...>]

On Thursday, 2010-02-25 at 12:13:00 +0100, Adrian Gajan wrote:

> Binding to TCP port 4949 on host 10.40.0.26                                                  
> 2010/02/25-09:13:14 Can't connect to TCP port 4949 on 10.40.0.26 [Cannot assign requested address]
>   at line 88 in file /usr/share/perl5/Net/Server/Proto/TCP.pm                                     

At this point something has already taken port 4949. Check the running
processes. Is it possible that you have two startscripts for munin in
/etc/rc?.d/S* ?

HTH,
Lupe Christoph
-- 
| There is no substitute for bad design except worse design.                   |
| /me                                                                          |

Re: [munin-users] Cannot assign requested address

[Adrian G. <ad...@ad...>]

On Thursday, 2010-02-25 at 12:13:00 +0100, Adrian Gajan wrote:

> Binding to TCP port 4949 on host 10.40.0.26                                                  
> 2010/02/25-09:13:14 Can't connect to TCP port 4949 on 10.40.0.26 [Cannot assign requested address]
>   at line 88 in file /usr/share/perl5/Net/Server/Proto/TCP.pm                                     

At this point something has already taken port 4949. Check the running
processes. Is it possible that you have two startscripts for munin in
/etc/rc?.d/S* ?

HTH,
Lupe Christoph
-- 
| There is no substitute for bad design except worse design.                   |
| /me                                                                          |

Hi, 

there is for each runlevel one startup / kill script:
mail:/home/adrian# find /etc/rc*.d -name "*munin*"
/etc/rc0.d/K20munin-node
/etc/rc1.d/K20munin-node
/etc/rc2.d/S98munin-node
/etc/rc3.d/S98munin-node
/etc/rc4.d/S98munin-node
/etc/rc5.d/S98munin-node
/etc/rc6.d/K20munin-node

I rebooted the box once more just for check if there some process have been already bind to 4949. Nothing.

mail:/home/adrian# netstat -tupan | grep 4949
mail:/home/adrian#
mail:/home/adrian# ping 10.40.0.1
PING 10.40.0.1 (10.40.0.1) 56(84) bytes of data.
64 bytes from 10.40.0.1: icmp_seq=1 ttl=64 time=0.194 ms
64 bytes from 10.40.0.1: icmp_seq=2 ttl=64 time=0.216 ms

The VPN gateway is accessible all the time.

Restarting munin-node by hand shows:

mail:/home/adrian# /etc/init.d/munin-node restart
Stopping Munin-Node: stopped beforehand.
Starting Munin-Node: done.

After a manual restart it works again.

Regards,

Adrian