| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 688.1 | MCS kit should be available shortly... | BSYBEE::EGOLF | John C. Egolf LKG2-2/T02 x226-7874 | Tue Feb 05 1991 18:26 | 18 |
| Bjorn,
There were a few performance related issues that found their
way into the product between the IFT kit and the EFT kit, but
they weren't the kind of things I would expect you to see
during a demo. They were the kinds of things that showed up
during actual performance testing.
Anyway, we have fixed these problems and are back to were we
were. We expect to announce our 'official' MCS (30 code freeze
kit) within a matter of days. Please keep your eye on this
notes conference for the announcement. I would check around
with your configuration of DECmcc ran well with one release,
but not the next.
Regards,
JCE
|
| 688.2 | I'll follow-up, too. | OSLACT::BJORN | Once again | Wed Feb 06 1991 02:25 | 9 |
| Thank you very much.
The problems I see can be if I use MCC> SHOW DOMAIN * or
MCC> SHOW NODE4 * ALL ATT or any other command that does a db-lookup.
The disk(s) are working heavily to find the data I requested. I have
not looked into these problems in detail, I've only used VPA, which of
course won't find anything unless the problem lasts for more than 2 minutes.
Bj�rn Olav Haugom
|
| 688.3 | Node4 access is much better. | TOOK::CAREY | | Wed Feb 06 1991 11:35 | 27 |
|
The EFT kit of V1.1 included significant additional functions in DECnet
Phase IV. We introduced the capability to work transparently with
single management link systems (such as DECrouters), and a consistency
check to ensure that management functions weren't being done to
Clusters instead of specific physical nodes. While these provided
significant flexibility, they introduced significant added link
overhead to our request management.
The new kit (announced today) has major improvements related to these
problem areas. There is a reduction of overhead for ALL Node4 requests
through some streamlining of our requests. And we added a small
caching capability to the Access Module so that repeated access to a
specified node yields significant improvement. After the first access
to a node4, we perform at speeds that often exceed the performance of
NCP for similar requests. Even the first access takes much less time
than it did with the EFT kit, without sacrificing flexibility or
consistency.
Additional improvements in our management of DECrouters has speeded
that up significantly as well, yielding additional perceived
capability.
Try the new kit, I think you'll like it.
-Jim Carey
|
| 688.4 | see note 3.49 for new kit | GOSTE::CALLANDER | | Wed Feb 06 1991 14:03 | 7 |
|
note 3.49 has the kit John was referring to. We are definitely
interested in hearing of any performance problems that still exist
with this latest kit.
thanks for the testing feedback
|
| 688.5 | DECmcc V1.1 performance info. requested. | ISIDRO::BURGOS | Luis Burgos. TIMG Spain. | Tue Mar 26 1991 07:40 | 7 |
| Could anybody provide performance figures for DECmcc ?.
I need to size a system to handle over 1000 alarms/events per
minute/second???. I know there are a lot of things involved in it
(Memory, I/O, Parameter setting, and so on...) But any performance data
would help a lot. If necessary contact me off-line.
Thanks.
|