[cups.general] On some machines CUPS does not start

Till Kamppeter till.kamppeter at gmx.net
Thu Jun 10 19:27:57 PDT 2004


Oi,

this problem was observed with both CUPS 1.1.19 and 1.1.21rc1.

In a network with many machines, all running CUPS and Mandrakelinux 9.2 
is a certain amount of boxes where the CUPS daemon does not start on 
boot time and claims that port 631 is already in use. One could think 
that this is caused by a firewall or by a bad order of starting the init 
scripts for the services on the machine. But there is no firewall 
installed and the machines where CUPS is working have the same order of 
starting of the image scripts. To investigate the problem more, a 
co-worker of mine has added a netstat call to the init script of CUPS, 
right before the start of the CUPS daemon, to see whether port 631 is 
occupied on any of the systems interfaces. But this was not true, 
netstat always did not report any use of port 631, but the CUPS daemon 
directly started afterwards stops due to a used port 631. After booting 
one can start CUPS normally on these machines. CUPS also sometimes 
crashes on these machines.

More you can read in this report

http://www.cups.org/str.php?L758

and there are also error_log files.

    Till





More information about the cups mailing list