Re: long checkpoints

How many disks? Are they on a RAID array? If so what level? Unless you have an exceptionally large number of separate disks or arrays I'd say your value for CLEANERS looks way too high. If you're using JABOD then unless they've radically changed the way these things are handled in the last couple of years, you could well have up to 50 processes randomly thrashing a disk, which would greatly increase seek times. If you're on RAID 5 then unless it's one of the new massively-cached models then this wouldn't surprise me either. Andy logicschmotom@yahoo.com (tomL) wrote in message news:<e70357eb.0306050753.39189be1@posting.google.com>... > hello all, > we have recently experienced checkpoints that are lasting 20-30 > seconds that were previously 5-8 seconds. > I am looking at the CONFIG file to see if something could be improved. > We are on HP-UNIX box (11.00)(2 proc box) with Informix 9.21 HC-5. > > BUFFERS 250000 # Maximum number of shared buffers > NUMAIOVPS 2 # Number of IO vps > PHYSBUFF 32 # Physical log buffer size (Kbytes) > LOGBUFF 32 # Logical log buffer size (Kbytes)> LOGSMAX 40 # Maximum number of logical log files > CLEANERS 50 # Number of buffer cleaner processes > LRUS 50 # Number of LRU queues > LRU_MAX_DIRTY 30 # LRU percent dirty begin cleaning> limit > LRU_MIN_DIRTY 25 # LRU percent dirty end cleaning limit > CKPTINTVL 300 # Check point interval (in sec) > MULTIPROCESSOR 1 # 0 for single-processor, 1 for> multi-processor > NUMCPUVPS 2 # Number of user (cpu) vps > SINGLE_CPU_VP 0 # If non-zero, limit number of cpu vps> to one > LTXHWM 50 # Long transaction high water mark> percentage > LTXEHWM 60 # Long transaction high water mark > (exclusive)> > KAIO is enabled. > > thanks for any help......