Re: Long Checkpoints

Ulrich Eckhardt wrote: > Hi, > > i have a problem with long checkpoint runs on an informix 7.30 UC 10 > running on linux. (Machine is a PIII 500 with 512 MB Memory). > > Sometimes checkpoints takes 15 to 25 seconds. During this > times i have approximately 2000 to 2300 dirty LRU queues. > > an onstat -FR (with only 205 dirty, but Fg Writes and LRU Writes > are also 0 with 2000 dirty). > ... > One thing wich i will try is to decrease the physical log and change > the LRU max-min parameters to halve the size. Reduce your LRU max and min (10/5, then 5/2, then 2/1), but do not fiddle with the size of your physical log. While the filling of the physical log (to 75%) does trigger a checkpoint, its more of a "panic checkpoint" rather than a "normal checkpoint". Which means, if your physical log is triggering a checkpoint under normal circumstances, its too small. > > But writing back 2000 queues makes approximately 4 MB of data. Could > this really take 20 seconds to write back ? or can i tune some > other parameters to get here a better performance ?. What about your dbspaces? Do you have more than 1 data dbspace on the same physical disk? That could cause I/O contention during checkpoints when multiple cleaners write simultaneously. Try and monitor I/O during a checkpoint (HP's glance is great! Is there some equivalent in Linux). Cheers Rudy