CUPS 1.3.6 - performance issue in AIX

angelb angelb at bugarin.us
Thu Mar 20 07:04:01 PDT 2008


> > Hello all.
> >
> > I got a chance to install and started testing CUPS 1.3.6:
> >
> > First, here's my observation in 1.3.5:
> > 1. When the client is set to "LogLevel debug2", it would take a very
> >    long time(>15min) for an lpstat to list 2700 printer queues or for
> >    an lpr(>9min) to print.
> >
> > 2. When the client is set to "LogLevel info", running an "lpstat -p",
> >    the cupsd daemon would die leaving no trace. The error_log file
> >    have no entry for it(possibly because it's in "info").
> >
> > 3. When the client is set to whatever LogLevel, AND, the remote.cache
> >    already exist, response time from lpstat(~5sec) to list all 2700
> >    printer queues or lpr(~3sec) to print.
> >
>
> Hi!
>
> I assume this is not AIX specific. We saw similar performance problems on Linux and HP-UX. But on our systems the cups daemon doesn't crash.
>

Our CUPS servers are running in Linux systems. Our clients are all AIX
and Solaris multi-processors and huge-gig of RAMS. We have true 2700+
printer queues shared between these clients. At CUPS 1.2, these clients runs cupsd very happily.

CUPS 1.3 brings new features, major and minor patches for performance,
security, and etc. But it also brings with it not seen in 1.2, at least
from what I've seen so far, issues that I've yet to convinced myself
that it's just my systems and not CUPS. BUT, how could I be confident to
say that it's my systems that are the issues if they run 1.2 just fine?

Anyway, I've yet to test CUPS 1.3.6 in Solaris. But I'm almost certain
the issues I saw in 1.3.5 will still be there...much like the ones in
AIX.

Hey, if there were no problems....I'd be out of a job, with or w/o CUPS!






More information about the cups mailing list