[cups.development] [RFE] STR #3692: Scheduler performance monitoring graph or graphs?

John A. Murdie john at cs.york.ac.uk
Fri Oct 15 09:47:02 PDT 2010


DO NOT REPLY TO THIS MESSAGE.  INSTEAD, POST ANY RESPONSES TO THE LINK BELOW.

[STR New]

Recently, my site had a situation where a Windows print server which was
configured to drive our CUPS scheduler by means of the Berkeley LPR
protocol and CUPS' cups-lpd(8) daemon caused an overload. The Windows
print server had several queues whose corresponding CUPS printers had been
deleted, but users continued to submit jobs to the Windows queues. CUPS
rejected each job, but the default action of the Windows print server was
to retry immediately. Suddenly, CUPS printing became very, very, slow -
ten minutes to print something - yet it wasn't immediately apparent to me
what the cause of the slowness was as none of the spurious jobs ever got
into a CUPS queue. Even watching error_log in DEBUG mode didn't provide an
immediately obvious clue, as it was so difficult to separate out the
crucial events.

Request for enhancement - provide some means of displaying the current
load on the CUPS scheduler, perhaps in historical and graphic format,
which would also enable the CUPS administrator to see how well the current
configuration was coping with the demands made upon it over time. Perhaps a
web page using SVG graphics to draw graphs in real-time?

Link: http://www.cups.org/str.php?L3692
Version:  -feature





More information about the cups mailing list