And that -too much fragmentation- is exactly why dedicating (part of) a 
volume to DUMP is really recommended.  This way you never will  encounter 
too much fragmentation.

Dedicating is not by using allocation type DUMP (as it was in VM/SP and 
HPO) but by alloacting as SPOL and adding DUMP (or something alike) to the 
CP_OWNED definition of the volume.  We also have VMUTIL check regularly if 
DUMP is still set to DASD, if not, we get an e-mail.

Kris,
IBM Belgium, VM customer support
> Actually, I think the fact since your power failure your spool has been
> higher tells me that before the power failure you were running with DUMP
> OFF.

> This can happen without your knowledge (want to see the scars?). The
> initial allocation of spool on 440RES is enough to get you started but
> not enough to keep running in production.

> What probably happened is this - at some time you IPLed the system
> (maybe even the first time you IPLed z/VM) and CP tried to allocate a
> dump space on spool. There wasn't enough room because the space had been
> fragmented too much to hold the contiguous space needed for the dump. So
> the system gave an error message and carried on.

> Eventually you cleaned up some space on spool and then the power failure
> happened and caused another IPL. This time CP found enough contiguous
> space and allocated the space. And your spool space suddenly became much
> fuller.

> The currently allocated space is as cleaned up as it can get - it is
> empty, but reserved for a future dump.

Reply via email to