I observed some details when I started to test with a Rexx server using WAKEUP (EXT. It appears that the SFS server sends the CMS client one interrupt telling "something has changed in something you know about" It is then up to the CMS client to decide when it will ask the SFS server what was changed. As long as the CMS client does not not send that question, no other interrupts are sent to him. As long as no SFS objects are touched, the CMS client doesn't have to bother the SFS server with his question "please send me the changes"
2009/4/2 Alan Altmark <alan_altm...@us.ibm.com> > > On Wednesday, 04/01/2009 at 09:46 EDT, Alan Ackerman > <alan.acker...@earthlink.net> wrote: > > > I was under the impression (from something I read here) that the cached > C > > MS list of files was > > updated when someone changed files on the directory. For SFS to inform > CM > > S of this would > > required that it know that CMS had it accessed. If this isn't what > happen > > s, I'll have to try to find > > that previous discussion and attempt to understand it again. > > This was addressed in another post. Yes, there is a cache of meta data. > > > If I access a file-control directory and issue LISTFILE * * fm, then > some > > one adds or deletes a file > > on the directory and I issue LISTFILE * * fm again, does it show the > diff > > erence? (I think it does.) > > Yes. > > > If > > so, how? Does it reload the entire file list each time from the SFS > serve > > r? If so, what is the value of > > caching the list? > > If memory serves, the cache allowed CMS to detect that someone else had > changed the file under you so that XEDIT, for example, could warn you that > the file had changed and "Are you sure?" (I have oh-so-vague > recollections that the cache was a late add.) \ > > > Same question for a dircontrol directory. (I suspect it doesn't show the > > difference.) > > The whole point of dircontrol directories is the access-to-release > consistency of the content. As long as you have the directory accessed, > no changes to the files in it will be visible to you, so it doesn't really > matter if the cache is present or not. > > Alan Altmark > z/VM Development > IBM Endicott -- Kris Buelens, IBM Belgium, VM customer support
