> VM will re-ipl when it doesn't have enough spool space.

Not necessarily.

I believe that you will find that if you separate page&swap from spool, and
make sure that there is adequate page&swap space, then the norm will be that
if the spool fills up, then the virtual spooled devices such as readers,
printers, etc of the guest virtual machines will be disabled by cp and cp
will continue to dispatch those VM's not constrained by (waiting on) spooled
devices and less frequently will abort. My memory tells me that the devices
become 'interrupt pending' or some such, its been a very long time since we
had spool mixed with paging devices.

The frequent cause of the abend is that CP wishes to page out one of his own
pages, asks for a page slot (either for page in or out), the page manager
has none (because page/swap/spool are mixed and spool ate all of the slots),
cp says bye because he cannot continue under these conditions.

The size of the spool is not as critical of a calculation as is the size for
page/swap for this reason.

In this case, when operations or the help desk informs you that there are
folks having troubles, then you can add spool space and the system should
not abort.

Note however, that if you have VM's that are key to your operation such as
VTAM, TCPIP, etc and they depend upon the spool for console logs or other,
then they may have major problems that can cause you to have a very bad day
even if you manage to add spool.

Regardless, I would recommend that if you wish to sleep well, split
page/swap from spool.

----------------------------------------------------------------------
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

Reply via email to