Thanks, Martha.  Running DB2 for VM with plenty of other load is the main

exposure, as the behavior which triggers the problem (contact me directly
 if
you want the gory details) is quite unusual and unique to DB2, as far as 
we
know.  There's a small possibility of hitting it while running a Linux th
at
exploits the Diagnose x'10' page release "ballooning" device driver hack,

but I doubt many folks run with that (I'm not even sure of its proper nam
e
or whether it ever made it into any distro).  Once it starts happening, o
f
course, the hang behavior starts spreading to other users, as the DB2
machines (or user machines trying to talk to it) almost invariably hold s
ome
other serialization when they hit the hang.  We'll be watching customer
experience with the circumvention as well as doing ongoing investigation 
to
determine whether it's an appropriate permanent solution (I hope it is, a
s
the alternative fixes are a lot more complicated).  The only concern we h
ave
with the circumvention isn't functional, it's the potential impact on
performance of holding a share of one of the locks involved (VMDPTIL) for
 a
longer duration.  

- Bill Holder, z/VM CP Development, Storage Management

Reply via email to