Hi, john.

Yes - I tried upgrading two weeks ago on an E450, 512MB, 2 cpu, loads of
swap.  Same thing.  Have also duplicated it on our test rig, which is a
sparc 1000 (sun 4d).  That will teach me to use a copy of the live dsm.opt
file as well as a copy of the db when I'm testing!

Also, if you just try going to v3.7.0, if you are using stk's acsls to
control your robotics, you may well have problems here as well, although
this version does not appear to have the max memory problem (its about 94MB,
btw - set buffsize to this, and no user sessions will be allowed to
start!!).

I raised a pmr on this.  When I last chased it yesterday 6apr00, the
response was that Tucson's testers had re-created the fault.

I must say that having previously tried to upgrade from 3.1.2.1 to v.30, and
v.40 before attempting 3.7.2 and failed each time with different problems,
mostly in testing, I find myself somewhat underwhelmed by IBM/Tivoli quality
control

regards,

dave