Re: Checkpoint is so sloooowww

andy@pneu.com offerred: +Hello !We use OnLine 5.08 on an digital-alpha server with 1 GB Ram and 4 +CPU.- Buffers 10000, Locks 40000, Physbuf 256, unbuffered Logging. - 9 +cleaners, 10 LRUS, clean beginns if 30% full, ends if 20% full. - checkpoint +300 sec. - root, log, phys and data in raw-devices on seperate Hdd's.Our +problem is: DB needs about 10 sec for an complete chkpoint! While checkpoint +is *running* all application-programms waiting forcompleting the checkpoint +and the users can't work.How can i reduce this 10 secs for the checkpoints? Lower LRU_MAX_DIRTY and LRU_MIN_DIRTY. The lower these values, the less work that needs to be done at checkpoint time. Right now you will be flushing up to 3000 buffers during teh checkpoint. I'd also set CLEANERS to the same value as LRUS to allow efficient LRU cleaning. You should also take a look at your i/o throughput (try sar or iostat) as 10 seconds for even 3000 buffers seems rather slow to me. -- Dave Kosenko davek@summitdata.com Director of Training Services (732) 469-4070 Summit Data Group (an Informix Authorized Education Center)