Hi, we have an 11.70 C5 Informix database of 90 GB and since 3 weeks we have encoutered some probelems, too simply the database is hanging , when we launch from the server
dbaccess mydbname its hanging and the database want to let us access and same thing for the clients.
the server is a Widnows 2008 r2, we have change the server, and also the same problem at random, we have started to analyze and we have seen thats in the online log at the momement of hanging it was a long checkpoints duration.
the time consuming is in flushing to disk, we have lower the LRUS to 2% and 1% but always flushing diration is high at random and also we have found 3 times crit_time is high (200 s)
we have 400 000 bufferpool and LRUS=20 and cleaners is 8.
The machine is a multiprocessor (4 cpu of 4 cores) with 53 GB of RAM.
for the reason why the checkpoints are occuring?
Look for it in the "trigger" - column.
Furthermore, onstat -g ckp would sometimes tell you the reason why a checkpoint is taking too long.
Most of the times this is due to a small physical log. The default value of PHYSFILE after initialisation is way to small for productive use.
So maybe you have to create a separate dbspace for your physlog and relocate it via onparams.