Good day,

We have problem with full checkpoints on Informix DB.
Sometimes checkpoints are more then 6 sec,somtetimes 50 sec.

Please help me to reslove problem.
Please see in attch. oconfig file and output from onstat -F command,and also interesting.

I'm guessing at some things since the information wasn't provided, but here's some observations:

You have 128 LRU's (good) but 128 CLEANERS (bad, since you only have 4 CPUVPS threads defined. I'm assuing 4 physical CPU's, so those cleaners are all fighting.

It looks like the majority of the "writes" occure on dbspace rtlogidxa. Cleaning that will be your bottleneck.

You are not running Fuzzy checkpoints - why not?

You have allowed AIO to take the default of 2*#chunks, yet you don't use AIO, so you have idle AIO threads just messing things up.

Physically, where does your physical and logical logs sit relative to rtlogidxa and rtlogdatac.

Thanks for the answer, the system has 2XDual Core CPU not 4 physical.

Fuzzy checpoints is not good for this application, it is specific.

Maybe somethink also I can modified in the system or onconfig file?

are you on a UNIX system. what does the storage configuration look like.
run iostat -xn 30 . let's see where the bottleneck is.

What we need is the onstat -c output.