Thank you Steve, this is the problem the phone call was about, but i am
sorry i miss number 4 in the original posting so, the actual configuration
we left last night is
4,5,6,7,8,9,10,12 for the memory cards and
11,13,14,15 for the processors.
which is basically the same you propose but with the difference that slot 12
should have a proccesor and slot 11 should have a memory card, ok i agree
but why is it that the manual A1809-90001, that is in a cd MPE/ix
Documentation collection that i have in page 60 states :
<<begin extract>>
997/T600 Only
Processors must be installed in the following sequence: PMB slots 15, 14,
13, 11, 10, and 9.
NOTE Multiprocessor systems have special power requirements. For more
information, refer to “Verifying Installation of Power Modules”.
<<end extract>>
also a label in the front PMB Card Cage skips slot 12 in the loading
sequence, but anyway like you say it probably make not make a difference,
the main problem is that the performance at this moment is to bad, we have
been looking at glance the whole day and the third bar from the top thas
shows memory utilization shows no 5 % utilization, also the message is
always MEMORY DEMAND IS HIGH, why is that if we have 6.1 GB of memory to
give ?...
This is getting me crazy mostly because all users are constantly complaining
about the slowness of the system..
We posted a report to Mexico´s HP response center but they have not give us
a good answer...
a hint of what could be happening will be welcome.. (Steve if you want me to
call you just send me an email)..
Thank you Again
Juan Hernández
Chihuahua México..
-----Original Message-----
From: Steve Cole [mailto:[log in to unmask]]
Sent: Friday, March 16, 2001 3:25 PM
To: [log in to unmask]; [log in to unmask]
Subject: Re: Re: [HP3000-L] Performance Problem
Juan,
The following is extracted from the "HP3000 997 Configuration Guide":
<< begin extract >>
The front PMB card cage contains 6 slots (10 through 15) for installation of
memory and
processor cards. The first processor is installed in slot 15.
The rear PMB card cage contains 9 slots (0 through 9) which are available
for memory
array cards, HP-HSC bus converters or processor cards. Note: The first
non-Large
Memory Subsystem array card is installed in slot 4 with add-on array cards
installed
in slots 5 through 11.
The following maximum PMB cards are supported:
six processor cards
eight memory array cards
six HP-HSC bus converter cards
<<end extract>>
According to your email:
> i looked into the PMB card cage and i guess we
> left a hole in the memory card loading sequence, the loading order we have
> is 6,8,9,12,7,5,3 for the memory cards and
> 15,14,13,11 for the processor cards, so as you can see we missed slot 10,
in
> the memory order.(i got the order from the 99x family instalation Guide).
Though slots 1 through 6 are available to memory, slot 4 should be used for
the first memory card. The first HP-HSC bus converter should be installed
in
slot 0.
Per the configuration guide you should probably move the processor in
slot 11 to slot 12 and the memory in slot 12 to slot 10. The memory in slot
3
should be moved to slot 4. (These changes may not make a difference
but would conform to the recommendations documented in the configuration
guide.)
Hope this helps,
Steve
=================================
Steve Cole
Outer Banks Solutions, Inc.
[log in to unmask]
www.outerbankssolutions.com
Phone: 919.231.2171 Fax: 919.231.7077
Sales: 800.368.6214
=================================
----- Original Message -----
From: Juan Hernandez <[log in to unmask]>
To: <[log in to unmask]>
Sent: Thursday, March 15, 2001 11:07 PM
Subject: Re: [HP3000-L] Performance Problem
> Thank you Mark, tried to run cstm but it stands forever and does not
> respond, on the other hand i looked into the PMB card cage and i guess we
> left a hole in the memory card loading sequence, the loading order we have
> is 6,8,9,12,7,5,3 for the memory cards and
> 15,14,13,11 for the processor cards, so as you can see we missed slot 10,
in
> the memory order.(i got the order from the 99x family instalation Guide).
>
> We have HP support and guess what, HP engineer did not notice about that.
>
> We will change memory card in slot 3 to slot 10 and see what happens.
>
> Do you think this can make a difference ?
>
> Thank you
>
> Juan
>
>
> -----Original Message-----
> From: Mark Bixby [mailto:[log in to unmask]]
> Sent: Thursday, March 15, 2001 3:30 PM
> To: [log in to unmask]
> Cc: [log in to unmask]
> Subject: Re: Performance Problem
>
>
>
>
> Juan Hernandez wrote:
> >
> > Thank you Mark, but how do i look at the memory error logs ?
>
> Wow, that cstm/mstm stuff is sure unfriendly (another word for "different
> than
> what I am used to on 6.0 and earlier") on 6.5. I just flailed around and
> ran
> this stuff for the first time on 6.5. I feel everybody's pain who has
> complained about this in the past on HP3000-L. ;-)
>
> What I did to view the memory log info:
>
> :cstm
> map (obtain the memory dev num)
> select device nnn (where nnn is the dev num from above)
> info (to start the tool running)
> wait (to wait for the tool to complete)
> infolog (to display the tool results)
> done
> exit
> ok
>
> - Mark B.
>
> >
> > -----Original Message-----
> > From: Mark Bixby [mailto:[log in to unmask]]
> > Sent: Thursday, March 15, 2001 1:53 PM
> > To: Juan Hernandez
> > Cc: [log in to unmask]
> > Subject: Re: Performance Problem
> >
> > Juan Hernandez wrote:
> > > Does anybody knows what a diaglogd process is running and why is
> it
> > > that is taking average 50 % of cpu (shows in glance).
> > > We have a 997-800 running 6.5 pp1, and just upgraded to 6 GB of
> > > memory but the performance is not good, the only weird thing we
noticed
> is
> > > that diaglogd process getting all the cpu...
> >
> > My recollection is hazy, but something in the back of my mind says that
> bad
> > memory on a 3000 or 9000 can make the diagnostics go crazy. Seeing as
how
> > you
> > just added a bunch of new memory, checking the memory error logs is
> probably
> > a
> > prudent idea.
> > --
> > [log in to unmask]
> > Remainder of .sig suppressed to conserve scarce California electrons...
>
> --
> [log in to unmask]
> Remainder of .sig suppressed to conserve scarce California electrons...
|
