Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Who will provide the ram?

0 views
Skip to first unread message

Nachman Yaakov Ziskind

unread,
Nov 2, 2003, 1:09:11 PM11/2/03
to sco...@xenitec.on.ca
... I just installed a 256mb SIMM into a Compaq Proliant, to add to the 384mb
that was already there. The Compaq Diagnostics show the new memory (640mb? 655?
can't remember) but the operating system (SCO OSE 506) seems to have problems
recognizing it, sometimes:

From /var/adm/messages:
mem: total = 392828k, kernel = 52916k, user = 339912k

From 'hw -r':

BaseMem: 638 Kb
ExtMem: 624 Mb

which looks good.

BootStr: hd(40)unix
Kernel: /unix -> /stand/unix

Available Memory

0x00000000-0x00007fff 32 Kb Kernel reserved
0x00008000-0x0009efff 604 Kb
0x0009f000-0x000fffff 388 Kb Non-RAM
0x00100000-0x00d90fff 12.57 Mb
0x00d91000-0x00ffffff 2.43 Mb Kernel data
0x01000000-0x17e11fff 366.07 Mb Shadow ram & other special mem
Non-ISA-DMAable
0x17e12000-0x17ffffff 1.93 Mb Kernel text
Shadow ram & other special mem
Non-ISA-DMAable
---------
383.62 Mb RAM total (384 Mb - 388 Kb)
which look less good.

and, from 'top':

Memory: 384M phys, 333M max, 310M free, 328M locked, 887M unlocked, 576M swap

So, how do I get the most out of my new memory?

Thanks!

--
_________________________________________
Nachman Yaakov Ziskind, EA, LLM aw...@egps.com
Attorney and Counselor-at-Law http://ziskind.us
Economic Group Pension Services http://egps.com
Actuaries and Employee Benefit Consultants

Jeff Liebermann

unread,
Nov 4, 2003, 3:19:19 AM11/4/03
to
On Sun, 2 Nov 2003 18:09:11 GMT, Nachman Yaakov Ziskind
<aw...@egps.com> wrote:

>... I just installed a 256mb SIMM into a Compaq Proliant, to add to the 384mb
>that was already there. The Compaq Diagnostics show the new memory (640mb? 655?
>can't remember) but the operating system (SCO OSE 506) seems to have problems
>recognizing it, sometimes:
>
>From /var/adm/messages:
>mem: total = 392828k, kernel = 52916k, user = 339912k

(...)

Also try running "memsize".
http://stage.caldera.com/cgi-bin/ssl_getmanpage?memsize+ADM+OS5

Do you have anything strange in /etc/default/boot in defbootstr?

What model Proliant? ECC ram? Mixed ECC and conventional? That goes
through self test on most models but duz weird things when the OS
loads. I've only tried it with W2K server which spent it's time
crashing until I replaced the RAM.

Sacrificing a ram in place of Isaac seemed to work for Abraham. You
might wanna babeque some old RAM and see if it helps.


--
# Jeff Liebermann 150 Felker St #D Santa Cruz CA 95060
# 831.336.2558 voice http://www.LearnByDestroying.com
# je...@comix.santa-cruz.ca.us
# 831.421.6491 digital_pager je...@cruzio.com AE6KS

Bela Lubkin

unread,
Nov 4, 2003, 3:54:41 AM11/4/03
to sco...@xenitec.ca
Nachman Yaakov Ziskind wrote:

> ... I just installed a 256mb SIMM into a Compaq Proliant, to add to the 384mb
> that was already there. The Compaq Diagnostics show the new memory (640mb? 655?
> can't remember) but the operating system (SCO OSE 506) seems to have problems
> recognizing it, sometimes:
>
> From /var/adm/messages:
> mem: total = 392828k, kernel = 52916k, user = 339912k

That's short...

> From 'hw -r':
>
> BaseMem: 638 Kb
> ExtMem: 624 Mb
>
> which looks good.

I believe those numbers come straight out of the BIOS, while these...

> BootStr: hd(40)unix
> Kernel: /unix -> /stand/unix
>
> Available Memory
>
> 0x00000000-0x00007fff 32 Kb Kernel reserved
> 0x00008000-0x0009efff 604 Kb
> 0x0009f000-0x000fffff 388 Kb Non-RAM
> 0x00100000-0x00d90fff 12.57 Mb
> 0x00d91000-0x00ffffff 2.43 Mb Kernel data
> 0x01000000-0x17e11fff 366.07 Mb Shadow ram & other special mem
> Non-ISA-DMAable
> 0x17e12000-0x17ffffff 1.93 Mb Kernel text
> Shadow ram & other special mem
> Non-ISA-DMAable
> ---------
> 383.62 Mb RAM total (384 Mb - 388 Kb)
> which look less good.

... are determined by more detailed examination of the system during
startup. You posted the "Available Memory" section of `hw -r mem`
output. Post the "Memory used by Kernel" section. This shows parts of
memory labeled "BIOS private area - Reserved". I'm guessing you'll find
your new 256MB labeled that way.

So then the question is, why?

Reboot and at the boot prompt, enter:

Boot
: mem=/v

The boot program will give you a detailed memory map. The areas shown
as "BIOS private area - Reserved" in `hw -r mem` should be shown in much
more detail. What does boot say about them?

Unfortunately I don't think it has a specific way of communicating what
the BIOS is probably trying to tell you, which is probably: "this memory
is bad or misconfigured". Run hardware diagnostics.

>Bela<

Nachman Yaakov Ziskind

unread,
Nov 4, 2003, 8:08:54 AM11/4/03
to


Oh, well:

Memory used by Kernel

0x00000000-0x00007fff 32 Kb Kernel reserved

0x00008000-0x0009fbff 607 Kb Unused
0x0009fc00-0x0009ffff 1 Kb BIOS private area - Reserved
0x000a0000-0x000effff 320 Kb Unused
0x000f0000-0x000fffff 64 Kb BIOS private area - Reserved
0x00100000-0x00d90fff 12.57 Mb Unused


0x00d91000-0x00ffffff 2.43 Mb Kernel data

0x01000000-0x17e11fff 366.07 Mb Unused


0x17e12000-0x17ffffff 1.93 Mb Kernel text
Shadow ram & other special mem
Non-ISA-DMAable

0x18000000-0xfebfffff 3692 Mb Unused
0xfec00000-0xfec0ffff 64 Kb BIOS private area - Reserved
0xfec10000-0xfedfffff 1.94 Mb Unused
0xfee00000-0xfee0ffff 64 Kb BIOS private area - Reserved
0xfee10000-0xfff7ffff 17.44 Mb Unused
0xfff80000-0xfffffffe 512 Kb BIOS private area - Reserved
---------
5.08 Mb RAM total (6 Mb - 939.001 Kb)

I *don't really understand the math here. It seems like all the numbers add up
to 5.08 mb.

But, no big chunk under BIOS private area - Reserved~

'3692 Mb Unused' - eh?



> So then the question is, why?
>
> Reboot and at the boot prompt, enter:
>
> Boot
> : mem=/v
>
> The boot program will give you a detailed memory map. The areas shown
> as "BIOS private area - Reserved" in `hw -r mem` should be shown in much
> more detail. What does boot say about them?
>
> Unfortunately I don't think it has a specific way of communicating what
> the BIOS is probably trying to tell you, which is probably: "this memory
> is bad or misconfigured". Run hardware diagnostics.
>
> >Bela<

Should I still do mem/v ? Can;t do it until tonight.

Thanks, Bela.

Nachman Yaakov Ziskind

unread,
Nov 4, 2003, 8:13:30 AM11/4/03
to je...@comix.santa-cruz.ca.us
Jeff Liebermann wrote (on Tue, Nov 04, 2003 at 08:19:19AM +0000):
> On Sun, 2 Nov 2003 18:09:11 GMT, Nachman Yaakov Ziskind
> <aw...@egps.com> wrote:
>
> >... I just installed a 256mb SIMM into a Compaq Proliant, to add to the
384mb that was already there. The Compaq Diagnostics show the new memory
(640mb? 655? can't remember) but the operating system (SCO OSE 506) seems to
have problems recognizing it, sometimes:
> >
> >From /var/adm/messages:
> >mem: total = 392828k, kernel = 52916k, user = 339912k
> (...)
>
> Also try running "memsize".
> http://stage.caldera.com/cgi-bin/ssl_getmanpage?memsize+ADM+OS5
>
> Do you have anything strange in /etc/default/boot in defbootstr?

Nope, all standard.



> What model Proliant? ECC ram? Mixed ECC and conventional?

1600. The only things I know about the RAM is a) it's advertised as
suitable for the Proliant (even printed on the package) and b) it fits
nicely in the slot and snaps in. And, oh yeah, it goes through the
Compaq diagnostics nicely.

> That goes
> through self test on most models but duz weird things when the OS
> loads. I've only tried it with W2K server which spent it's time
> crashing until I replaced the RAM.
>
> Sacrificing a ram in place of Isaac seemed to work for Abraham. You
> might wanna babeque some old RAM and see if it helps.

Chanukah coming. Maybe I should lubricate the server with olive oil?

Jeff Liebermann

unread,
Nov 4, 2003, 1:21:46 PM11/4/03
to
On 4 Nov 2003 08:13:30 -0500, Nachman Yaakov Ziskind <aw...@egps.com>
wrote:

>1600. The only things I know about the RAM is a) it's advertised as

>suitable for the Proliant (even printed on the package) and b) it fits
>nicely in the slot and snaps in. And, oh yeah, it goes through the
>Compaq diagnostics nicely.

Single or Dual processor? Which processor?
6/266 6/350 6/400 6/550

The 1660 and 1600R are good machines but somewhat picky about the
memory quality and matching. I had to juggle the ram among several
other servers to get a combination that worked consistantly. However,
mine failed the Compaq diagnostics. It seemed to want fairly
identical SDRAM but it's been a while and I don't recall the results
of my RAM juggling experiments.

Did you perhaps mix ECC RAM with non-ECC RAM? That will pass self
test but do weird things.

Scarey suggestion. Remove the known working RAM and leave the memory
you added. If it totally trashes your filesystem, it's bad RAM.
However, I don't think even this drastic test will help as I suspect
it's a side effect of mixing conventional and ECC RAM. Make sure you
have a good backup ready.



>> Sacrificing a ram in place of Isaac seemed to work for Abraham. You
>> might wanna babeque some old RAM and see if it helps.

>Chanukah coming. Maybe I should lubricate the server with olive oil?

Olive Oyl is not much of a sacrifice. You could comb the neighborhood
looking for the traditional ram with horns stuck in a thicket, but
finding a thicket seems as difficult as finding some working Compaq
RAM. Every time I get into trouble, I barbeque an old 286 or 386
motherboard on the hibachi. There's a minimum of toxic fumes and the
subsequent divine intervention into my affairs seems to be quite
helpful.


--

Jeff Liebermann 150 Felker St #D Santa Cruz CA 95060

(831)421-6491 pgr (831)336-2558 home
http://www.LearnByDestroying.com AE6KS
je...@comix.santa-cruz.ca.us je...@cruzio.com

Mike Brown

unread,
Nov 4, 2003, 3:16:18 PM11/4/03
to
Jeff Liebermann wrote:
>
> On 4 Nov 2003 08:13:30 -0500, Nachman Yaakov Ziskind <aw...@egps.com>
> wrote:
>
> >1600. The only things I know about the RAM is a) it's advertised as
> >suitable for the Proliant (even printed on the package) and b) it fits
> >nicely in the slot and snaps in. And, oh yeah, it goes through the
> >Compaq diagnostics nicely.
>
> Single or Dual processor? Which processor?
> 6/266 6/350 6/400 6/550
>
> The 1660 and 1600R are good machines but somewhat picky about the
> memory quality and matching. I had to juggle the ram among several
> other servers to get a combination that worked consistantly. However,
> mine failed the Compaq diagnostics. It seemed to want fairly
> identical SDRAM but it's been a while and I don't recall the results
> of my RAM juggling experiments.
>
> Did you perhaps mix ECC RAM with non-ECC RAM? That will pass self
> test but do weird things.
>

I don't think the 1600 will accept non-ECC ram, the BIOS should display
a non-compatible memory error at boot. On the 350/400/450 generation
the memory was PC100 ECC SDRAM.


> Scarey suggestion. Remove the known working RAM and leave the memory
> you added. If it totally trashes your filesystem, it's bad RAM.
> However, I don't think even this drastic test will help as I suspect
> it's a side effect of mixing conventional and ECC RAM. Make sure you
> have a good backup ready.

What does happen if you boot on the new ram only? With ECC ram the worst
that should occur is the HW detects a non-correctable memory error
and halts the system.


>
> >> Sacrificing a ram in place of Isaac seemed to work for Abraham. You
> >> might wanna babeque some old RAM and see if it helps.
>
> >Chanukah coming. Maybe I should lubricate the server with olive oil?
>
> Olive Oyl is not much of a sacrifice. You could comb the neighborhood
> looking for the traditional ram with horns stuck in a thicket, but
> finding a thicket seems as difficult as finding some working Compaq
> RAM. Every time I get into trouble, I barbeque an old 286 or 386
> motherboard on the hibachi. There's a minimum of toxic fumes and the
> subsequent divine intervention into my affairs seems to be quite
> helpful.
>
> --
> Jeff Liebermann 150 Felker St #D Santa Cruz CA 95060
> (831)421-6491 pgr (831)336-2558 home
> http://www.LearnByDestroying.com AE6KS
> je...@comix.santa-cruz.ca.us je...@cruzio.com


At least I now know who to check with for computer pagan tips.

Mike

--
Michael Brown

The Kingsway Group

Bela Lubkin

unread,
Nov 4, 2003, 4:41:46 PM11/4/03
to sco...@xenitec.ca
Nachman Yaakov Ziskind wrote:

> Memory used by Kernel
>
> --> 0x00000000-0x00007fff 32 Kb Kernel reserved
> 0x00008000-0x0009fbff 607 Kb Unused
> --> 0x0009fc00-0x0009ffff 1 Kb BIOS private area - Reserved
> 0x000a0000-0x000effff 320 Kb Unused
> --> 0x000f0000-0x000fffff 64 Kb BIOS private area - Reserved
> 0x00100000-0x00d90fff 12.57 Mb Unused
> --> 0x00d91000-0x00ffffff 2.43 Mb Kernel data
> 0x01000000-0x17e11fff 366.07 Mb Unused
> --> 0x17e12000-0x17ffffff 1.93 Mb Kernel text


> Shadow ram & other special mem
> Non-ISA-DMAable
> 0x18000000-0xfebfffff 3692 Mb Unused

> --> 0xfec00000-0xfec0ffff 64 Kb BIOS private area - Reserved
> 0xfec10000-0xfedfffff 1.94 Mb Unused
> --> 0xfee00000-0xfee0ffff 64 Kb BIOS private area - Reserved
> 0xfee10000-0xfff7ffff 17.44 Mb Unused
> --> 0xfff80000-0xfffffffe 512 Kb BIOS private area - Reserved


> ---------
> 5.08 Mb RAM total (6 Mb - 939.001 Kb)
>
> I *don't really understand the math here. It seems like all the numbers add up
> to 5.08 mb.

5.08Mb is the sum of the lines I marked "-->" above. It's a silly
number -- sum of memory used by the initial loading of the kernel +
memory which was marked by the BIOS as unusable. It's important for the
kernel to know about areas marked as BIOS-private, so it doesn't try to
find and allocate regular RAM there, but it makes no sense to lump the
two things together into a single sum. Then it's additionally confusing
to throw these all out in a single table, interleaved with the unused
areas, and show a summation below which only sums some of the lines!

You'll notice that the BIOS has marked out various areas of different
sorts. 9FC00-9FFFF is regular RAM being used as private storage by the
BIOS. F0000-FFFFF is BIOS ROM, likely overlaid with faster RAM in a
modern system, but that RAM will have been made read-only by the BIOS,
using the CPU's MTRRs (Memory Type and Range Registers).
FEC0xxxx are addresses used by an I/O APIC; FEE0xxxx are the local APIC;
FFF80000-FFFFFFFF I'm not sure.

Most of these are areas that will not contain RAM due to the inherent
design of the CPU and PC architecture. So it's silly for `hw` to report
these as areas the kernel is "using", as if it chose to allocate them.

> But, no big chunk under BIOS private area - Reserved~

Right. This means that booting with "mem=/v" probably won't show
anything, but you might want to do it anyway just to be sure.

> '3692 Mb Unused' - eh?

This particular map is showing the entire 4GB 32-bit address space, with
usage information describing what the memory map looked like at kernel
startup time. Because it's an address space map (rather than a RAM
map), "Unused" areas can represent present-but-unallocated RAM or
not-present RAM, which is what the 3692MB area is.

> Should I still do mem/v ? Can;t do it until tonight.

Up to you. I suspect it will not add anything except to give you
slightly more detail on the BIOS private areas we already see. But you
should be rebooting anyway so you can run more intensive diagnostics...

>Bela<

Nachman Yaakov Ziskind

unread,
Nov 5, 2003, 11:21:46 AM11/5/03
to
Bela Lubkin wrote (on Tue, Nov 04, 2003 at 09:41:46PM +0000):
> Nachman Yaakov Ziskind wrote:
> >
> > Should I still do mem/v ? Can;t do it until tonight.
>
> Up to you. I suspect it will not add anything except to give you
> slightly more detail on the BIOS private areas we already see. But you
> should be rebooting anyway so you can run more intensive diagnostics...
>
> >Bela<

Thank you for that enlightening technical discourse.

Unfortunately(?) the ram (marked 256MB SYNC 100 MHZ ECC from SimpleTech,
by the way, and sold for use in the Proliant) flunked the Compaq diagnostics.

So, back to square one. I'm wondering if the ram is defective, or just
unsuitable for the 1600.

Nachman Yaakov Ziskind

unread,
Nov 5, 2003, 11:25:48 AM11/5/03
to
Mike Brown wrote (on Tue, Nov 04, 2003 at 08:16:18PM +0000):
> Jeff Liebermann wrote:
> >
> > On 4 Nov 2003 08:13:30 -0500, Nachman Yaakov Ziskind <aw...@egps.com>
> > wrote:
> >
> > >1600. The only things I know about the RAM is a) it's advertised as
> > >suitable for the Proliant (even printed on the package) and b) it fits
> > >nicely in the slot and snaps in. And, oh yeah, it goes through the
> > >Compaq diagnostics nicely.

Not the detailed ones. Just the preliminaries.

> > Single or Dual processor? Which processor?
> > 6/266 6/350 6/400 6/550

P3-600 single.

> > The 1660 and 1600R are good machines but somewhat picky about the
> > memory quality and matching. I had to juggle the ram among several
> > other servers to get a combination that worked consistantly. However,
> > mine failed the Compaq diagnostics. It seemed to want fairly
> > identical SDRAM but it's been a while and I don't recall the results
> > of my RAM juggling experiments.
> >
> > Did you perhaps mix ECC RAM with non-ECC RAM? That will pass self
> > test but do weird things.
> >

Simple Tech 256MB SYNC 100MHZ ECC. Marketed for the Proliant (unsure if I
specified 1600 P3-600).



> I don't think the 1600 will accept non-ECC ram, the BIOS should display
> a non-compatible memory error at boot. On the 350/400/450 generation
> the memory was PC100 ECC SDRAM.
>
>
> > Scarey suggestion. Remove the known working RAM and leave the memory
> > you added. If it totally trashes your filesystem, it's bad RAM.
> > However, I don't think even this drastic test will help as I suspect
> > it's a side effect of mixing conventional and ECC RAM. Make sure you
> > have a good backup ready.
>
> What does happen if you boot on the new ram only? With ECC ram the worst
> that should occur is the HW detects a non-correctable memory error
> and halts the system.


Will not boot (disk/tape drives whir, no screen, no beep codes).

As I mentioned in my reply to Bela, ram flunks Compaq diagnostics
consistently.

Thanks for your input!

Bob Bailin

unread,
Nov 5, 2003, 12:21:45 PM11/5/03
to

"Nachman Yaakov Ziskind" <aw...@egps.com> wrote in message
news:2003110511...@egps.egps.com...

> Mike Brown wrote (on Tue, Nov 04, 2003 at 08:16:18PM +0000):
> > Jeff Liebermann wrote:
> > >
> > > On 4 Nov 2003 08:13:30 -0500, Nachman Yaakov Ziskind <aw...@egps.com>
> > > wrote:
> > >
> > > >1600. The only things I know about the RAM is a) it's advertised as
> > > >suitable for the Proliant (even printed on the package) and b) it
fits
> > > >nicely in the slot and snaps in. And, oh yeah, it goes through the
> > > >Compaq diagnostics nicely.
>
> Not the detailed ones. Just the preliminaries.
>
> > > Single or Dual processor? Which processor?
> > > 6/266 6/350 6/400 6/550
>
> P3-600 single.
>
> > > The 1660 and 1600R are good machines but somewhat picky about the
> > > memory quality and matching. I had to juggle the ram among several
> > > other servers to get a combination that worked consistantly. However,
> > > mine failed the Compaq diagnostics. It seemed to want fairly
> > > identical SDRAM but it's been a while and I don't recall the results
> > > of my RAM juggling experiments.
> > >
> > > Did you perhaps mix ECC RAM with non-ECC RAM? That will pass self
> > > test but do weird things.
> > >
>
> Simple Tech 256MB SYNC 100MHZ ECC. Marketed for the Proliant (unsure if I
> specified 1600 P3-600).

Please check out the memory search utility at www.kingston.com
You'll see that there are dozens of Compaq with the Proliant brand name, and
they use several different versions of 256MB memory modules. Even if
you specified the Proliant 1600, the 266MHz submodel uses a different
version than the other 1600 models such as your 6/600.

You can order direct from Kingston on their website, if you wish.

Bela Lubkin

unread,
Nov 5, 2003, 12:36:11 PM11/5/03
to sco...@xenitec.ca
Nachman Yaakov Ziskind wrote:

> > > >1600. The only things I know about the RAM is a) it's advertised as
> > > >suitable for the Proliant (even printed on the package) and b) it fits
> > > >nicely in the slot and snaps in. And, oh yeah, it goes through the
> > > >Compaq diagnostics nicely.
>
> Not the detailed ones. Just the preliminaries.

> Simple Tech 256MB SYNC 100MHZ ECC. Marketed for the Proliant (unsure if I
> specified 1600 P3-600).

> As I mentioned in my reply to Bela, ram flunks Compaq diagnostics
> consistently.

Hmmm. It's interesting that OpenServer's reaction to this was simply to
ignore the RAM. How did it know it was bad? I'd like you to go do that
"mem=/v" now that you have the machine in pieces on the floor... one
part of the output should be an "E820" report, which is the BIOS's best
guess at what memory the OS ought to use. Does it include or exclude
the bad 256MB?

If E820 excludes the bad memory, the BIOS knew it was bad and should
have said something to you...

If E820 includes it, OpenServer /boot's memory test simply failed to
detect that the RAM existed. It tests for the existence of memory on
4KB boundaries, and the test isn't particularly sophisticated, so I
would guess it must be the BIOS.

>Bela<

Nachman Yaakov Ziskind

unread,
Nov 5, 2003, 12:48:41 PM11/5/03
to
Bela Lubkin wrote (on Wed, Nov 05, 2003 at 05:36:11PM +0000):
> Nachman Yaakov Ziskind wrote:
>
> > > > >1600. The only things I know about the RAM is a) it's advertised as
> > > > >suitable for the Proliant (even printed on the package) and b) it fits
> > > > >nicely in the slot and snaps in. And, oh yeah, it goes through the
> > > > >Compaq diagnostics nicely.
> >
> > Not the detailed ones. Just the preliminaries.
>
> > Simple Tech 256MB SYNC 100MHZ ECC. Marketed for the Proliant (unsure if I
> > specified 1600 P3-600).
>
> > As I mentioned in my reply to Bela, ram flunks Compaq diagnostics
> > consistently.
>
> Hmmm. It's interesting that OpenServer's reaction to this was simply to
> ignore the RAM. How did it know it was bad? I'd like you to go do that
> "mem=/v" now that you have the machine in pieces on the floor...

"In pieces on the floor." Heh, heh. Production server, (I'm writing this
email on it!) - mem was yanked last night when it tested bad. I will try
to schedule some more downtime tonight and test it further.

> one
> part of the output should be an "E820" report, which is the BIOS's best
> guess at what memory the OS ought to use. Does it include or exclude
> the bad 256MB?
>
> If E820 excludes the bad memory, the BIOS knew it was bad and should
> have said something to you...
>
> If E820 includes it, OpenServer /boot's memory test simply failed to
> detect that the RAM existed. It tests for the existence of memory on
> 4KB boundaries, and the test isn't particularly sophisticated, so I
> would guess it must be the BIOS.
>
> >Bela<

--

Bill Vermillion

unread,
Nov 5, 2003, 1:15:02 PM11/5/03
to
In article <Jkaqb.19751$Ig3....@newssvr32.news.prodigy.com>,

And BTW the tech support at Kingston is quite good as they seem to
know about problems in specific HW. When RAM for a Cicso didn't
work, I called and they gave me the exact part number of the ROM
on the Cicso router I needed to upgrade, and the the minimum
BIOS and IOS needed.

You normally don't think of tech support on memory but they were
far better than 95% of the tech support I've used.

Recently when I called a vendor for some SCSI stuff I told them
exactly what I needed and was told 'call back after 9AM PDT and
talk to tech support'.

I did, and when I explained the problem to the sales who answered
and was transfered to tech support when I stated the problem
the tech support person said [and I quote exactly] "I have no idea
what you are taling about". He put me to another tech, who then
passed me off to a 3rd. I got an instant answer and two
solutions.

But if I had not been sure of what I was speaking of and been a
typical end users or HW support with a PC mindset the solution to
the customer would be a HW and OS upgrade [at least $2000 + labor].

I solved for a phone call on their nickel, 1 hour of my time, and
a CC charge of $199.

But I was really impressed with Kingston and I've never had a
problem with Cicso. Oh that I could say that for other vendors.
[The only problem with Cisco was a 4AM EST call and I wound up in
a European support and the accent was a bit tough to translate :-)]

Bill

--
Bill Vermillion - bv @ wjv . com

Bela Lubkin

unread,
Nov 5, 2003, 2:30:54 PM11/5/03
to sco...@xenitec.ca
Nachman Yaakov Ziskind wrote:

> > Hmmm. It's interesting that OpenServer's reaction to this was simply to
> > ignore the RAM. How did it know it was bad? I'd like you to go do that
> > "mem=/v" now that you have the machine in pieces on the floor...
>
> "In pieces on the floor." Heh, heh. Production server, (I'm writing this
> email on it!) - mem was yanked last night when it tested bad. I will try
> to schedule some more downtime tonight and test it further.

It's really just idle curiousity on my part -- don't take down a
production server to test it...

>Bela<

0 new messages