Rob van der Heij <[EMAIL PROTECTED]> wrote: >When the FST on the CMS disk is written, the timestamp in the >directory is updated. Even when you just access a R/W disk, CMS >already writes back the FST (that's what triggers the non-shared S/Y >stat in the SYSPROF). You can use the mdiskblk stage to get block 3 >and then 4 or 5 for the date. >I can code an example if you need that. Used to have a pipeline stage >that would re-access the disk and pull unseen records from the PROP >logging (like tail -f in Linux).
A minor correction (Rob knows this, he stated it correctly in a later post, but I want to clarify): The timestamp in the directory is updated by RELEASE, not ACCESS. So if you accidentally ACCESS a disk R/O, you can do one of three things: 1) RELEASE it, updating the timestamp; 2) DETACH (or reDEFINE) it, thus moving it before CMS can update the timestamp and causing an implicit RELEASE (due to the machine check); 3) ACCESSing another disk "over" it -- for example, if you have 0190 R/W as C, "ACCESS 19E C". (2) and (3) above both avoid the timestamp update (OK, to be 100% correct, I haven't verified that (3) does these days, but I'd be surprised if CMS is fancy enough to do otherwise). ...phsiii