I kept the LPARs but I really cut down on the OSA adapters. We have 4 OSA cards.
2 that supports the ICC devices 2 that supports are normal network traffic When I defined all the addresses and then gave all addresses to all 4 lpars. Now, all the addresses on the DS6800 are on all LPARs, but I also cut back to what we would realistically use, not the max that we could define for the DS6800. Tom Duerbusch THD Consulting >>> [EMAIL PROTECTED] 8/16/2006 4:14 PM >>> Did you cut back your HSA by removing LPARs, reducing MAXDEVS, or both? Mostly I want to verify that MAXDEVS impacts HSA size. Brian Nielsen On Wed, 16 Aug 2006 15:33:43 -0500, Tom Duerbusch <[EMAIL PROTECTED]> wrote: >I do believe that it is the max devices defined, but that is also across >LPARs. > >What initially bit me, was I defined 4 lpars. Why? I don't know. I >needed two, a 390 and an IFL. But I thought about a test one and >perhaps a "systems" one. And the convention was to define the devices >to all LPARs. Well, almost 2 GB for HSA convinced me otherwise. Even >if I didn't bring up the LPAR, the definitions took their pound of HSA, >hence real storage. > >I cut things back to the minimum of 768M HSA size. > >I remember when OS/VS 2 took 768K for the superviser. Is the HSA >process 1,000 times more complex then OS/VS 2? > >Tom Duerbusch >THD Consulting > >>>> [EMAIL PROTECTED] 8/16/2006 2:32 PM >>> >On a z/890, does the size of the HSA change if the value of the MAXDEVS > >parameter on the RESOURCE statement in the IOCP is increased/decreased? > >The z/890 System Oveview manuals says the HSA size varies according to > >the "size and complexity of the I/O configuration", but it's not clear >if >it refers to the actual defined devices or the maximum devices that can >be >defined. It makes the most sense for it to be the maximum devices, but > >I'd like to verify it. > >Brian Nielsen >=========================================================================