CUPS 1.2rc1 - CPU load

[angelb at bugarin.us]

Hi all.

I've downloaded and installed CUPS 1.2rc1 and loaded it with 1167
printer queues. That's the number of printers we currently support
under 1.1.23.

I've noticed that when starting CUPS on the server, CPU load increases
as much as 100%. See the following:

ps -ef|grep cupsd (every five seconds):  Nearing FULL reload completion
. 
. 
root      2938     1 99 10:08 ?        00:08:39 /usr/sbin/cupsd
root      2938     1 98 10:08 ?        00:08:39 /usr/sbin/cupsd
root      2938     1 97 10:08 ?        00:08:39 /usr/sbin/cupsd
root      2938     1 96 10:08 ?        00:08:39 /usr/sbin/cupsd
root      2938     1 95 10:08 ?        00:08:39 /usr/sbin/cupsd
root      2938     1 94 10:08 ?        00:08:39 /usr/sbin/cupsd
root      2938     1 93 10:08 ?        00:08:39 /usr/sbin/cupsd
root      2938     1 92 10:08 ?        00:08:39 /usr/sbin/cupsd
root      2938     1 91 10:08 ?        00:08:39 /usr/sbin/cupsd
root      2938     1 91 10:08 ?        00:08:39 /usr/sbin/cupsd
root      2938     1 90 10:08 ?        00:08:39 /usr/sbin/cupsd
root      2938     1 89 10:08 ?        00:08:39 /usr/sbin/cupsd
root      2938     1 88 10:08 ?        00:08:39 /usr/sbin/cupsd
root      2938     1 88 10:08 ?        00:08:39 /usr/sbin/cupsd
...


Running vmstat the same time as ps.
vmstat 5: Nearing FULL reload completion
. 
. 
procs -----------memory---------- ---swap-- -----io---- --system-- ----cpu----
...
 1  0    160   2996 172384 225200    0    0     0    24 1020    24 100  0  0  0
 1  0    160   2916 172384 225200    0    0     0    44 1030    31 100  0  0  0
 1  0    160   2924 172384 225200    0    0     0    25 1018    26 100  0  0  0
 1  0    160   2852 172384 225200    0    0     0    22 1024    26 100  0  0  0
 1  0    160   2788 172384 225200    0    0     0    24 1019    24 100  0  0  0
 0  0    160   2908 172384 225204    0    0     0    34 1023    33 49  0 50  1
 0  0    160   2908 172384 225204    0    0     0    33 1021    69  0  0 99  1
 0  0    160   2884 172384 225204    0    0     0    21 1023    29  0  0 99  0
 0  0    160   2876 172384 225204    0    0     0    38 1024    27  0  0 99  0
procs -----------memory---------- ---swap-- -----io---- --system-- ----cpu----
 r  b   swpd   free   buff  cache   si   so    bi    bo   in    cs us sy id wa
 1  0    160   2892 172384 225204    0    0     0    22 1021    28  0  0 99  0
 0  0    160   2820 172384 225204    0    0     0    20 1018    23  0  0 99  0
 0  0    160   2820 172384 225204    0    0     0    20 1021    27  0  0 99  0
 0  0    160   2748 172384 225204    0    0     0    20 1018    23  0  0 99  0
 0  0    160   2724 172384 225204    0    0     0    30 1024    32  1  0 99  1


localhost$ lpstat -v|wc
   1167    4668   46776

Systmes info:
localhost$ cat redhat-release
Red Hat Enterprise Linux ES release 4 (Nahant)

localhost$ cat meminfo
MemTotal:       515716 kB

localhost$ cat cpuinfo
processor       : 0
vendor_id       : GenuineIntel
cpu family      : 6
model   : 11
model name      : Intel(R) Pentium(R) III CPU family      1400MHz
stepping        : 1
cpu MHz         : 1396.342
cache size      : 512 KB

I've also noticed that when the load is complete, high CPU use on
cupsd lingers on for awhile but slowly settles down.

I tested CUPS 1.2b2 on an AIX system, 2x451MHz and 6Gb RAM, and
noticed a simillar reaction where cupsd pegs the CPU.

Can someone please verify this on 1.2rc1? Create some 1500 printer
queues and stop and start cupsd several times and see if you have
simillar result?

Thanks!
Angel