|
Hi there,
Well you asked...
� What is the exact performance problem?
Basically, I'm trying to overcome my own stupidity...smile. I
*had* 5 domains setup for RECORDING of entities, I've grouped them by
types of entities, ie: DNA4_SYSTEMS, TERMINAL_patients, er SERVERS,
etc. I also setup 5 EXPORTER batch queues, using the same *logic*???
I've since spoken to one of the developers, thank you Sam...smile,
and decided to consolidate ALL historical data into one domain! I like
it. I'm in the process of propagating all the historian/recordings,
and then I'll setup the exporter again...
I was seeing alot of PAGE FAULTING, via MONITOR SYSTEM; I was
seeing close to 100% CPU utilization for looooooooong periods of time;
and I was seeing alot of RWMBX processes, mostly the batch stuff, and
the EXPORTER subprocesses... From, SHOW SYSTEM, I saw alot of swapped
out processes, mostly the UCX stuff, not doing much with my SNMP stuff,
YET!
The configuration is EMS V2.2, with VMS 5.5-1. Memory Managment is
*actually* working like a charm.... trolling and tickin' away...smile.
I've since then added 6 100k pagefiles, set all volumes NO_HIGHWATER,
installed MCC_MAIN.EXE, which I understand didn't buy me much, and said
alot of HAIL MARYs....smile.
I finally got up the nerve to ask the customer to consider a, gulp,
LARGER Server...smile. We running on a VAX3400 with 36mg memory, and a
Client of a DEC433 PC... I've been receiving alot of hand holding, and
great suggestions from the MCC team...!
If you've got more VMS tuning "magic", pray tell???
/Mic
|
| > If you've got more VMS tuning "magic", pray tell???
>
Well, I think Dav Lyons had a few hints tucked away in the MCC-TOOLS
conference. You may want to poke around there and look for pointers to
the "survival" kit. I think there are some system tuning swags in
there that may help.
|