Hi Josef,

>1)Because FLUSH command did not always work, we can not be sure that data
> are saved on hard disk
I personally haven't seen ***on the same machine*** a situation where the data was not "visibly changed" - the case of erik was special with the underlying view, and even there "my" schemes cut down the errors.

I have more than a passing interest in this topic, because we use table records for semaphores between n processes.

>2) Because sys(1023) and set reproces commands did not always work, we can
>not be sure if we see all data that other user saved on disk a minute ago.
Set reprocess ?? works only if you try to lock!
Try altering set refresh!!!!

hth

thomas