[cups.general] Jbase and cups and a stuck queue, hopefully someone recognizes this problem.

[Mark Krenz]

I think I've proved that this failure happens right after this overnight 
process runs.  So that's most likely the cause, but of what I still 
don't know.

 The overnight process finished at midnight so I tried printing a job 
larger than 17KB to cups and it failed.  Smaller jobs went through.  I 
tried restarting cups and it still had the same problem.  I tried even 
restarting the localhost interface and it still had the problem.  But at 
7:15am this morning when I got into the office I thought of trying to 
just restart nfs (unrelated) since that is something that runs in kernel 
space for the most part.  And lo and behold that fixed the printing 
problem this morning.  So we didn't have to restart the server today, 
just nfs.  So would restarting nfs be resetting something in kernel 
space that would fix the cups issue?

Thanks,
Mark

Mark Krenz wrote:
> Kurt Pfeifle wrote:
>   
>> Did you ever consider to re-build your CUPS RPM? And use 1.1.23 sources
>> instead of the ancient 1.1.17?
>>
>> Did you ever sift through the Release Notes listing all changes since the
>> 1.1.17 release (see http://www.cups.org/relnotes.php)? Didn't you see any-
>> thing that may be relevant to your case (fixed race conditions, certificate
>> creation bugs,...)?
>>   
>>     
> I have gone through the change log for cups and also glibc to see if any 
> bug was fixed.  Nothing.  As I mentioned before, we have another machine 
> practically identical in versions and function and it acts as a control 
> group because we haven't had the problem on that machine in the 2.5 
> years that it has been active.
>
> We've done extensive comparison between the two machines and the major 
> differences are that the versions of jbase are different, the glibc 
> version is at a different patch level (2.3.2-95.33 vs. 2.3.2-95.44).
>
> I would point the finger at jbase and say that it could definitely be 
> the trigger of the problem, but why would that cause it to fail between 
> lpr and cups?
>
>
>
>   


-- 
Mark S. Krenz
Linux System Administrator
Cook Incorporated
mark.krenz at cookmedical.com
phone: (812)339-2235 ext.4224
Cell: 322-8117
alt. cell: 322-4449