On Tuesday, 01/27/2009 at 12:06 EST, "Schuh, Richard" <rsc...@visa.com> wrote: > Isn't there a contradiction between the first and the last two? It > cannot be both free and have no unused cylinders or tracks. The F5 DSCBs > determine what is free; the F1, F3 and F6 define what is allocated. If > the first or only F5 is all zeros, there is no free extent.
I, too, was stymied by the seeming contradiction. First of all, the first F5 DSCB can only be empty if the VTOC is indexed. The index bit is off for a CPVOL, so the F5 DSCB must have valid stuff in it. I came to understand that the F5 really describes an *extent*, not "free space". I imagine this is related to the way MVS allocates space for datasets, whereby the allocation is done in some units (e.g. tracks) that result in an extent >= size of dataset. Available space (tracks/cyls) for that extent is kept in that extent description. So the F5 DSCB describes the remainder of the volume as a single extent and indicates that there are no cylinders or tracks available within the extent. Since you cannot create an additional extent (already have described the whole volume) and no space is available in that extent, the volume is full. I'm working to put all this on my home page so that future generations won't forget! :-) Alan Altmark z/VM Development IBM Endicott
