[cups.bugs] cups 1.1.21 socket backend problem?

David Schlenk david-schlenk at bethel.edu
Tue Oct 19 12:31:59 PDT 2004


First, brief overview of my setup:

2 backend cups servers that actually send the jobs to printers
3 frontend cups servers, 2 of which serve out the load balanced backend 
queues via netatalk (two different subnets), and the other servers them 
out via samba.

I upgraded to 1.1.21 a week ago or so and ever since have been 
experiencing a couple queues stop themselves because of a job that seems 
to get stuck in the queue. The printer becomes unavailable on port 9100, 
but I can still ping it from the print server.  The printer won't start 
printing again until I clear out the queue and restart the printer, and 
then cancel whatever job is stuck on the printer one last time (if you 
cancel it prior to clearing out the cups queue it keeps coming back). 
It seems like some kind of two way communication is breaking down, but I 
can't really tell from the logs what's going on.  There doesn't seem to 
be any pattern to what kind of jobs stop it, but I believe once it was a 
word file and at least twice it was a PDF file.

Another note - when one of my backend cups servers goes down, it seems 
to also crash the frontend cups server that runs on the samba box. 
Anyone else experience this? Its been happening to me since at least 
1.1.18. Kinda makes loadbalanced/failsafe rather useless...
-- 
David Schlenk
Operating Systems Analyst
Bethel University
david-schlenk at bethel.edu





More information about the cups mailing list