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 Historical content remains viewable.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

IBM ServeRAID: Array question

9 views
Skip to first unread message

Al Savage

unread,
Nov 9, 2001, 1:22:17 AM11/9/01
to
OK, all you ServeRAID gurus, lend me your brain.

I've got an original IBM ServeRAID (not II, Ultra, or 3x/4H series), the
older model with "only" three separate SCSI channels. At the moment,
I've got it installed in an IBM PC365 (6589-13U) that runs WSeB SMP from
it's own controller and HDD.

I've created the 4.70.17 ServeRAID CD and booted from it, updated the
ServeRAID's BIOS (2.23.1 -> 4.70.17) and firmware (97246 -> 2.25.01?).

Now I need to define a logical drive, but I just ran into something that
all my reading of the various IBM .txt files and the ServeRAID pdf
didn't prepare me for: it looks as if each array of drives must be on
one (ONE!) channel? Is this correct?

What I want to do is have at least one HDD in an external enclosure. Do
I need to have ALL the HDDs for that array in the external enclosure?

The docs are clear about one thing: using the (single, on this model)
external connector disables the internal channel 1 connector.

--
Regards,
Al S.

* Hillman & Rootes Group manuals online: http://asavage.fdns.net/Hillman
* Ford Falcon manuals online: http://FalconFAQ.fdns.net
This OS/2 system ("Tori", W4 FP15) uptime is 0 days 00:21 hours

Kenneth Taylor

unread,
Nov 9, 2001, 10:06:00 PM11/9/01
to
I have two of the original Serveraid cards, but never tried to have the
logical drives interleaved on different channels, but I believe that it
has
that capability, just like the AMI Megaraid cards do.

I have my drives in the server enclosure, not in an external cabinet.
Yes, using the external connector does disable the connector on channel
1 from being used. You can use an internal 68 pin to hd 50 or 68 pin
ribbon cable
to an external backplate connector and use one of the other channels to
drive the
external box though and use channel 1 for internal.

I had upgraded my bios version to at the time 4.50 and used the images
on
that version CD to make the bios update floppies and used them instead
of
the boot CD. Version 4.50 used Windows 98 as the host OS to do this and
it
never worked. I could not get it to recognize my combination of Symbios
controllers and the Serveraid.

I found that it was easier to use the older DOS version 3.50 setup
floppy,
which appears to not be available anymore from IBM's website to setup
and initialize the logical drives as the setup CD boot image did not
have drivers
for my combination of controllers and trying to fiddle with
uncompressing
the boot image on the CD to a floppy and modifying it with the
appropriate
DOS drivers just plain did not work. I finally modified the autoexec to
find the
appropriate CD drive letter but 98 would then cough up giving a IOS
error.

I have tried the newest version 4.80 CD using Linux, but it won't work
on either
of the systems the Serveraids are in either.

The original Serveraid is not fast, but it is like a rock and once you
get used to
the setup it' s really pretty easy to work with. The RAID manager
software is nice too.
I wish however that I could find out if the addon battery backup cache
card ever
was available and could get my hands on two of them. I see that the
model II card
had this and the model I has the same connector, but I have not found
any references
that even alludes to if the model II daughtercard will work with it or
not.


Al Savage

unread,
Nov 10, 2001, 1:25:27 AM11/10/01
to
On Sat, 10 Nov 2001 03:06:00, Kenneth Taylor <kta...@satx.rr.com>
wrote:

> I have two of the original Serveraid cards, but never tried to have the
> logical drives interleaved on different channels, but I believe that it
> has that capability, just like the AMI Megaraid cards do.

Hmmm. Using RaidMan, it refused to allow me to choose an array using a
drive on another channel. I thought that this would be easy (like
duplexing), but no.

> I have my drives in the server enclosure, not in an external cabinet.
> Yes, using the external connector does disable the connector on
> channel 1 from being used. You can use an internal 68 pin to hd
> 50 or 68 pin ribbon cable to an external backplate connector and
> use one of the other channels to drive the external box though
> and use channel 1 for internal.

I see in the PDF that that was actually an available IBM option kit, to
bring the channel 2 & 3 connectors out to the back. It appears easier
to do with the II and newer, they made the board layout fit the single
bracket more compactly; ours use a second slot bracket for that option.

I think I may have a custom cable made by CS Electronics
(http://www.scsi-cables.com/) , one that has two M68s (for the adapter,
say channel 2, and the other for one internal HDD) and an F68 panel
mount for the case back. Then I can run a cable to the external
enclosure and continue the array there.

It's a valid workaround, anyway. Or, I may just mount the whole array
external, I'm not decided yet. I have to do this x2.

> I had upgraded my bios version to at the time 4.50 and used the
> images on that version CD to make the bios update floppies and
> used them instead of the boot CD.

I did that at first, using
http://www-1.ibm.com/servlet/support/manager?rs=0&rt=0&org=psg&doc=MIGR-
4X7QVE

25P2552 IBM ServeRAID BIOS and Firmware Update Diskette 1 of 2 version
4.70
ftp://ftp.pc.ibm.com/pub/pccbbs/pc_servers/25p2552.exe

25P2553 IBM ServeRAID BIOS and Firmware Update Diskette 2 of 2 version
4.70
ftp://ftp.pc.ibm.com/pub/pccbbs/pc_servers/25p2553.exe

because I couldn't make the CD from the .iso (305MB!) here -- CDRecord
was willing, but AudioCDCreator wasn't -- but eventually I went to the
Win box and burned the CD there.

But: the BIOS update floppy appears to ONLY update the BIOS (in my case,
2.23.1 -> 4.70.17) but NOT the firmware! OTOH, only the first diskette
is actually bootable, and it autoruns the BIOS update. I couldn't
figure out what to do with the second diskette: it won't boot, and the
update program on the first diskette doesn't ask for it. I figured that
it contained BIOS images for other model ServeRAIDs, but that's just a
guess.

When you boot the CD, the first thing it does is check the BIOS and
firmware revisions, and update them automatically. I didn't expect it
to find anything to update the first time I booted the CD (because I'd
already run the update floppy), but it did: it then updated the firmware
only (in my case, from 97246 -> 2.25.01).

I don't know why the diskette BIOS update version does not (apparently)
update the firmware also.

> Version 4.50 used Windows 98 as the host OS to do this
> and it never worked.

What a surprise ;/

> I found that it was easier to use the older DOS version 3.50 setup
> floppy, which appears to not be available anymore from IBM's
> website

Right. All the .txt files say that you MUST now use the CD to define
the array(s) and logical drive(s) (not true: you can use the OS utility,
in our case ServeRAID Manager [x:\raidman\raidman.cmd] to do this).
What a pain. I didn't play around with it much before I updated the
ServeRAID's BIOS, but it appeared that on the old BIOS version, you
could define the array(s) and logical drive(s) right from the <CTRL-I>
BIOS screen, which makes more sense to me :)

> to setup and initialize the logical drives as the setup CD
> boot image did not have drivers for my combination of
> controllers and trying to fiddle with uncompressing
> the boot image on the CD to a floppy and modifying it
> with the appropriate DOS drivers just plain did not work.

I was able to find the newer IPSRAID.ADD on the CD OK, and copy it to
the boot drive, then add the appropriate line in CONFIG.SYS, but I'd
have thought that they'd have provided a utility to screw that up for
me! Well, I can add "BASEDEV=IPSRAID.ADD /V" as well as most folk, and
I know to put the .ADD and .DDP in x:\OS2\BOOT, so that went OK.

Interestingly, IPSRAID.ADD on the 4.70.17 CD is newer than the FP2
version for WSeB.

> I finally modified the autoexec to find the appropriate CD
> drive letter but 98 would then cough up giving a IOS
> error.

Good ol' W98 and IOS.

> I have tried the newest version 4.80

4.70.17 is the latest I could find; let me know where you found newer,
please.

> CD using Linux, but it won't work on either
> of the systems the Serveraids are in either.

Oh? It works quite well here on the 325. As it only modifies the
ServeRAID, it doesn't need to know about other controllers. It's kind
of slow to boot (on this dual P200 box), but it eventually got there.

> The original Serveraid is not fast, but it is like a rock and
> once you get used to the setup it' s really pretty easy
> to work with.

It seems "fast enough" to me, especially considering the cost of it
compared to the 3L or 4H ! Man, those things are *expensive*. Hard for
me to sell into my clients' offices.

> The RAID manager software is nice too.

It seems "OK" so far. It's a Java app, and I've yet to find a Java app
that I like.

Do you know if the ServeRAID Manager needs to be running in order for
the log to be updated, and for a RBL to occur? I ask because I haven't
done but one test RBL (removed one drive from the enclosure, stuck
another one in: ServeRAID Manager beeps when the bay goes to "defunct"
status, autodetects the new drive and began a rebuild without
intervention, though it beeped every three minutes while RBL'ing).

I'll probably look at IPSSEND.exe to detach the drive from the array
before removing it, in the future.

> I wish however that I could find out if the addon battery
> backup cache card ever was available

It's listed in the PDF; I just assumed that it was actually
manufactured.

> and could get my hands on two of them. I see that the
> model II card had this and the model I has the same
> connector, but I have not found any references
> that even alludes to if the model II daughtercard will work
> with it or not.

I run UPSs on all these boxes, big 'uns, so I wonder at the value of
having the cache backed up. I guess if the array went down hard, it
*might* be useful, but it's hard for me to imagine that scenario.

--
Regards,
Al S.

* Hillman & Rootes Group manuals online: http://asavage.fdns.net/Hillman
* Ford Falcon manuals online: http://FalconFAQ.fdns.net

This OS/2 system ("Tori", W4 FP15) uptime is 0 days 04:43 hours

Al Savage

unread,
Nov 10, 2001, 2:30:33 AM11/10/01
to
On Sat, 10 Nov 2001 06:25:27, asa...@iname.com (Al Savage) wrote:

> > I have tried the newest version 4.80
>
> 4.70.17 is the latest I could find; let me know where you found newer,
> please.

Got it:
http://www.pc.ibm.com/qtechinfo/MIGR-495PES.html?lang=en_US&page=brand&b
rand=root&doctype=&subtype=Cat

(line wrap will hurt you)

Downloading the .iso now: it's bloated up to 500MB! But the summary of
changes (over 4.70) does look enticing, including:

* A new Configuration wizard for ServeRAID configuration that provides
an easier-to-use interface, better access to status indicators, and
enhanced functionality that eliminates potential configurations that are
not valid.

* Added ability to set online physical drives to the defunct state.

Oh, it looks as if the 3.50C utilities are available from at link at the
bottom of the page. You said you were looking for them?

Poking around IBM's site, I find a note that
ServeRAID II UltraSCSI Adapter 6H3584* 76H3587 "OS/2 is not supported"

Oh, I see. Just for that one model of server, the 720. No such note
for the same controller in a, say, 325.

It looks like the ServeRAID II backup cache's FRU is
ServeRAID II 8MB/Battery Backup Cache
76H5401* 21H8973

But there's this note:
ServeRAID 3H 32 MB Battery Backup Cache
28L1003* 21H8979 footnote (4) (5)
(4) The Netfinity ServeRAID 3H 32MB Battery-Backup Cache is supported in
the Netfinity ServeRAID 3H Ultra2 SCSI Adapter (part number 01K7207) and
the IBM ServeRAID II Ultra SCSI Adapter (part number 76H3584)

(5) When installed in a ServeRAID II Ultra SCSI Adapter, only 4 MB of
the battery-backup cache memory is utilized. The firmware on this
adapter must be at level 2.86.03, or above

So, there's at least TWO battery backup caches for the II.

I like that the II costs $1900:
76H3584
http://www5.pc.ibm.com/us/products.nsf/$wwwPartNumLookup/_76H3584

Here's our old original:
ServeRAID PCI SCSI-2 Adapter
70G8489* 09N9541
replaces 76H6875
replaces 76H3574
replaces 06H9334
(Withdrawn October 30, 1997)

No, I can't find the FRU or Option for the original cache.

--
Regards,
Al S.

* Hillman & Rootes Group manuals online: http://asavage.fdns.net/Hillman
* Ford Falcon manuals online: http://FalconFAQ.fdns.net

This OS/2 system ("Tori", W4 FP15) uptime is 0 days 05:40 hours

Al Savage

unread,
Nov 10, 2001, 3:18:04 AM11/10/01
to
> No, I can't find the FRU or Option for the original cache.

Well,
ftp://ftp.pc.ibm.com/pub/pccbbs/pc_servers/09n1055.pdf
Page 275 (PDF page 289) shows the part numbers for the ServeRAID through
the 3H.

Parts listing
Part Name FRU No.
IBM PC ServeRAID Adapter 06H9334
IBM ServeRAID II Ultra SCSI Adapter 76H3587
IBM ServeRAID II Internal 3rd Channel Cable 01K6501
IBM ServeRAID II DASD Status Cable 01K6496
IBM Netfinity ServeRAID 3H Ultra2 SCSI Adapter 01K7396
IBM Netfinity ServeRAID 3HB Ultra2 SCSI Adapter 01K7396
IBM Netfinity ServeRAID 3H/3HB 3rd Channel Cable 01K7249
IBM Netfinity ServeRAID 3H/3HB/3L DASD Status Cable 03K9356
IBM Netfinity ServeRAID 3L Ultra2 SCSI Adapter 01K7207
32 MB IBM NVRAM Battery-Backup Option 21H8979
8MB IBM NVRAM Battery-Backup Option 21H8973

This implies that, perhaps, both of the Battery backup options will work
with the 3 and lower series. Maybe.

--
Regards,
Al S.

* Hillman & Rootes Group manuals online: http://asavage.fdns.net/Hillman
* Ford Falcon manuals online: http://FalconFAQ.fdns.net

This OS/2 system ("Tori", W4 FP15) uptime is 0 days 07:07 hours

Everyone

unread,
Nov 10, 2001, 11:06:39 AM11/10/01
to

Al Savage wrote:

> I like that the II costs $1900:
> 76H3584
> http://www5.pc.ibm.com/us/products.nsf/$wwwPartNumLookup/_76H3584

I saw this the other day:

http://www.compgeeks.com/details.asp?invtid=76H2668

ServeRAID I adapter for $29 US. Talk about depreciation.

Al Savage

unread,
Nov 10, 2001, 11:38:35 AM11/10/01
to

Oh, there's one at eBay right now for $10, no bids, and "Buy it Now" for
$19. I was thinking about picking up a second one, but I'll probably
fork out the $120 or so to buy a II.

--
Regards,
Al S.

* Hillman & Rootes Group manuals online: http://asavage.fdns.net/Hillman
* Ford Falcon manuals online: http://FalconFAQ.fdns.net

This OS/2 system ("Tori", W4 FP15) uptime is 0 days 15:39 hours

Kenneth Taylor

unread,
Nov 11, 2001, 12:16:37 AM11/11/01
to
Al,

The latest firmware for the Serveraid is 2.25.01.

The IBM motherboards that had it onboard have I think 2.88.01
as the latest firmware revision.

The BIOS is the higher number 4.50,4.70;etc.

The link I used to find the 4.80 is the one you found:
http://www.pc.ibm.com/qtechinfo/MIGR-495PES.html

Did you get the dumplog utilites?

I purchased mine originally from Compgeeks about 8 months ago.
They just got in another batch about a month ago and the price is much
lower than when I bought them. I believe I paid $79 at the time, still
a good price IMO.

At the time I bought mine I did not have any hot swappable cartridges,
I was just using SCA to 68 pin adapters. I have since bought 5 Antec
SCA2 hot swap removable cartridges (about $90-$99).

If the manager is running it will provide support for rebuild if you
have set
the BIOS to not automatically rebuild the array, but there is no daemon
that
gets installed and runs, so you will not get any log updates if the
manager is
not running, but I am unsure how many events are kept internally in the
cards
memory or potentially lost if it's not running.

You can resync up the array and other config maintenance thru the
manager
or the CL programs. I am a little mystified as to if IBM changed the
firmware
and BIOS on the originals to automatically sync the array. IBM mentions
a lot
of changes in 4.50-4.80 but I think it may just pertain to the newer
cards in the
series and not apply to the original unless it specifically states
"Serveraid" .

I think the reason that the BIOS no longer supports setting up the
logical drives
is that so many changes were made to it to accomodate the later series
cards that they
ran out of room to support it, with the original card as the lowest
common denominator
for eeprom size.


0 new messages