Hello all,

Our DB has noticeably gotten slower over the past few months [since
Art last gave me a "hey your DB is purring along fine"].  I checked,
and our Read Utilization is down around 95%, which is rather low --
the last check I did it was closer to 99.97%.

In reviewing the newsgroup, it seems that our BUFFERS may need some
tuning, as our DB is now closer in size to 12GB -- rather than 8GB
previously.  We're doing more, so I assume our "working set" is
greater.  Granted, we probably have some I/O issues, we're in the
process of solving, but in the meantime I can't seem to get the
BUFFERS above 350000.

I've set SHMTOT to 0, and SHMADD and SHMVIRTSIZE are 50502.  LOCKS is
700000.  As below:

LOCKS           700000          # Maximum number of locks
BUFFERS         350000          # Maximum number of shared buffers
NUMAIOVPS       20              # Number of IO vps
PHYSBUFF        48              # Physical log buffer size (Kbytes)
LOGBUFF         32              # Logical log buffer size (Kbytes)
LOGSMAX         10              # Maximum number of logical log files
CLEANERS        100             # Number of buffer cleaner processes
SHMBASE         0x10000000L     # Shared memory base address
#SHMVIRTSIZE    16384           # initial virtual shared memory
segment size
SHMVIRTSIZE     50502           # initial virtual shared memory
segment size
SHMADD          50502           # Size of new shared memory segments
(Kbytes)
SHMTOTAL        0               # Total shared memory (Kbytes).
0=>unlimited
CKPTINTVL       400             # Check point interval (in sec)
LRUS            80              # Number of LRU queues
LRU_MAX_DIRTY   30              # LRU percent dirty begin cleaning
limit
LRU_MIN_DIRTY   20              # LRU percent dirty end cleaning limit
#LRU_MAX_DIRTY  5               # LRU percent dirty begin cleaning
limit
#LRU_MIN_DIRTY  1               # LRU percent dirty end cleaning limit
LTXHWM          50              # Long transaction high water mark
percentage
LTXEHWM         60              # Long transaction high water mark
(exclusive)
TXTIMEOUT       300             # Transaction timeout (in sec)
LTXEHWM         60              # Long transaction high water mark
(exclusive)
TXTIMEOUT       300             # Transaction timeout (in sec)
STACKSIZE       32              # Stack size (Kbytes)

This is a dual athlon 1900 box, running 7.31.UD2, and RH7.3.  We've
got 3.5GB of RAM on the machine.  I've setup the shmmax and shmall to
be ~640MB, or 671088640.  This doesn't help much.  Anytime I try and
go over 350,000 BUFFER's, it quits immediately with:

23:29:09  Exception Recursion while handling: MT_EX_OS, Context: mem
23:29:09  PANIC: Attempting to bring system down
23:29:09  mt_shm_remove: WARNING: may not have removed all/correct
segments

Any ideas why?