Long Checkpoints IDS 11.50

Hi all, last Friday and yesterday we had checkpoints with duration of 290 and 135 seconds. All writing applications were blocked. Typically the duration of checkpoints is 0 to 2 seconds. Didn't I understand right that blocking checkpoints shouldn't occur any more in this release (IDS 11.50.FC5W4)? What could be the reason for such long blocking checkpoints? TIA, Reinhard.

On Dec 8, 10:22 am, "Habichtsberg, Reinhard" <RHabichtsb...@arz- emmendingen.de> wrote: > Hi all, > > last Friday and yesterday we had checkpoints with duration of 290 and 135 > seconds. All writing applications were blocked. Typically the duration of > checkpoints is 0 to 2 seconds. > > Didn't I understand right that blocking checkpoints shouldn't occur any more > in this release (IDS 11.50.FC5W4)? What could be the reason for such long > blocking checkpoints? > > TIA, > Reinhard. You would want to look at the onstat -g ckp output. Yes, most checkpoints should be the new style, interval checkpoints which should only block writing for a very small period (and not while the buffer pool is flushing). But certain activity can still trigger the old style, block while writing the buffer pool to disk, type checkpoints. Jacques Renaut IBM Informix advanced support