On Aug 12, 2004, at 2:03 PM, James Melin wrote:
1) V-DISK blocks appear to be 512 byte blocks & are created in 8-block
pages. True?
Dunno about the 8-block pages. But they are 512-byte blocks.
2) Will all Linux guests, that contain "MDISK 300 FB-512 V-DISK 256000
MWV" share the same V-DISK space? If so is one 256000 block area
allocated
& shared or (256000 x #-of-Linux's) allocated & all shared?
Each VDISK space is private to the guest. VDISK is not shareable.
3) How will you automatically know when a given Linux is the first and
therefore needs to format the area, if area(s) are shared?
You won't: each guest has its own space.
4) SuperValu Red Paper suggests V-DISK size 15% of VM size, what's up
with
this formula 2 x VM size + 32M ??
I don't remember the paper in any detail. If that's *all* the swap
they have, I'd be concerned that either they fail to allocate memory
pretty often, or that they have too much virtual storage defined for
their guests and they're wasting a lot of storage in file caching. If
that's high-priority swap backed by low-priority swap on real DASD it
sounds reasonable. If their applications cause very little fluctuation
in the amount of memory the system uses (for example, a select-based
(not fork-and-exec or thread-creating) server, fixed-size preallocated
buffers) then 15% might be adequate.
David's recommendation basically comes from the old-school Unix 2.5 x
physical memory rule-of-thumb; I usually start at 1-to-1.5 times
physical memory, assuming that the memory has been squeezed down enough
that very little is used for file or buffer cache.
Adam
----------------------------------------------------------------------
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390
