t5120 does not recognize disk

[pa...@trifox.com]

Used T5120, I put in a fresh disk drive. show /SYS/HDD0/PRSNT says that the drive is there, the green light is on, and yet

{ok} show-disks

does not show it, it shows only five entries (I am not in front of it now so I cannot show them).

{ok} probe-scsi-all

makes the disk light blink but show-disks still does not show it, so clearly I cannot install Solaris because there is no disk.

I thought, ok the brand new drive is broken. I found another T5120, went to check it out, same thing. Its two drives light up but {ok} show-disks never shows anything.

I must be doing some very basically wrong here because I figure it is not likely that two different systems with different drives exhibit the same show-disks behavior. I would like to get this going to give my Ultra 5(!) running Sol 10 a break and also because with ILOM, I can start/stop it remotely.

I presume that

{ok} show-disks

supposed to show the scsi drives? It dawned on me that maybe it shows only devices that have a bootable OS but that seems farfetched.
Thanks,
Paul

[DoN. Nichols]

On 2016-08-14, pa...@trifox.com <pa...@trifox.com> wrote:

> Used T5120,

I don't have a T5120, but my T5220 should be similar enough.

> I put in a fresh disk drive. show /SYS/HDD0/PRSNT says
> that the drive is there, the green light is on, and yet > > {ok}
> show-disks

> does not show it, it shows only five entries (I am not in front of it
> now so I cannot show them).

> {ok} probe-scsi-all

> makes the disk light blink but show-disks still does not show it, so
> clearly I cannot install Solaris because there is no disk.

> I thought, ok the brand new drive is broken. I found another T5120,
> went to check it out, same thing. Its two drives light up but {ok}
> show-disks never shows anything.

Do either of the systems have Solaris on them? Or are both of
them needing the OS installed?

O.K. First question: Is this a used disk?
If so -- what system was it in before?

Sun machines of the X series (AMD and Intel processors) use the
fdisk format instead of Sun's normal format.

Secondly -- if it was in a hardware RAID setup, it has a
different format applied to the disk.

Have you actually tried running format on it? Or even just
"format" to see whether it is among the disks listed?

In a *booted* solaris system, you need to create the device
entries. To do that -- once the disk is in place, type:

devfsadm -c disk

(and if there is a dead drive to replace from the same slot, make that):

devfsadm -c disk -C

The upper case 'C' scrubs away old entries which don't match any current
disk.

Once that is done, you can try running format to see the disk.

And once you see it -- select "partition" to see what the
partition table looks like. A normal disk will look something like
this:

======================================================================
Part Tag Flag Cylinders Size Blocks
0 root wm 0 - 29 130.96MB (30/0/0) 268200
1 swap wu 30 - 59 130.96MB (30/0/0) 268200
2 backup wu 0 - 65532 279.36GB (65533/0/0) 585865020
3 unassigned wm 0 0 (0/0/0) 0
4 unassigned wm 0 0 (0/0/0) 0
5 unassigned wm 0 0 (0/0/0) 0
6 usr wm 60 - 65532 279.11GB (65473/0/0) 585328620
7 unassigned wm 0 0 (0/0/0) 0
======================================================================

And one currently or previously in use in a RAID system, including zfs,
will look something like this:

======================================================================
Part Tag Flag First Sector Size Last Sector
0 usr wm 256 279.39GB 585921082
1 unassigned wm 0 0 0
2 unassigned wm 0 0 0
3 unassigned wm 0 0 0
4 unassigned wm 0 0 0
5 unassigned wm 0 0 0
6 unassigned wm 0 0 0
8 reserved wm 585921083 8.00MB 585937466
======================================================================

If it looks like the latter, (note the slice 8 at the end, and
the absence of the "whole disk" at slice 2) you need to run format again
-- but with the "-e" (expert mode) option, and once you have the disk in
format, opt to label it and it will ask you which of two formats to use.
Pick the first of the two offered.

And -- if it was previously in a system with AMD or Intel
processors, you will probably first need to select 'type' and let it
read the disk parameters from the disk (second thing offered -- the
first is to take the info from format.dat -- the second is what works
for me.)

BTW -- are these SAS or SATA drives? Both will work, but it is
expecting SAS.

> I must be doing some very basically wrong here because I figure it is
> not likely that two different systems with different drives exhibit the
> same show-disks behavior. I would like to get this going to give my
> Ultra 5(!) running Sol 10 a break and also because with ILOM, I can
> start/stop it remotely.

You need to get the drive configured with the standard SPARC
disk format, not the fdisk which the Intel/AMD systems use.

If you don't have a bootable Solaris on either of the T5120
systems, but you say you do have an Ultra-5 system, maybe you can
connect a drive docking box tot he Ultra-5 via USB and re-format through
there. But I would expect a very slow USB on that, so it could take
forever.

> I presume that
>
> {ok} show-disks

> supposed to show the scsi drives? It dawned on me that maybe it shows
> only devices that have a bootable OS but that seems farfetched.

Perhaps only devices which have a normal SPARC type format, and
not those which have Intel/AMD fdisk formats. I don't normally use
"show disks", but I do regularly use "probe-scsi-all". I've had
sufficiently weird formats on various systems. Some Fibre Channel (FC-AL)
disks which were in some unknown hardware RAID system had even been
formatted with an uncommon sector size (540 bytes instead of 512 bytes),
and I needed to go into the format subcommand and it took nearly forever
to re-format to 512 bytes/sector.

And I have also encountered some 146 GB FC-AL drives which had
been in some IBM system which had such a different sector size which I
have never been able to re-format properly.

I hope that this helps.

Good Luck,
DoN.

--
Remove oil spill source from e-mail
Email: <BPdnic...@d-and-d.com> | (KV4PH) Voice (all times): (703) 938-4564
(too) near Washington D.C. | http://www.d-and-d.com/dnichols/DoN.html
--- Black Holes are where God is dividing by zero ---

[pa...@trifox.com]

Hi DoN,
It's a brand new, never used drive: SAS Seagate Savio 300Gb. IIUC you are saying that it must be formatted by Solaris before Solaris can be installed on it. Seems a real "chicken-egg" problem or perhaps a way for Oracle to require users to purchase drives through them. When the installation fails (no disk found), it drops into a shell where I tried format but of course, there are no drives.

probe-scsi-all returns:

{0} ok probe-scsi-all
/pci@0/pci@0/pci@9/scsi@0
Waiting for AAC Controller to start: . . Started

AAC Kernel Version: 15583


/pci@0/pci@0/pci@8/pci@0/pci@9/SUNW,qlc@0,1
QLogic QLE2462 Host Adapter Driver(SPARC): 1.24 11/15/06
Firmware version 4.00.26
Fibre Channel Link down
Possible causes: No cable, incorrect connection mode or data rate

/pci@0/pci@0/pci@8/pci@0/pci@9/SUNW,qlc@0
QLogic QLE2462 Host Adapter Driver(SPARC): 1.24 11/15/06
Firmware version 4.00.26
Fibre Channel Link down
Possible causes: No cable, incorrect connection mode or data rate

/pci@0/pci@0/pci@2/scsi@0

MPT Version 1.05, Firmware Version 1.22.00.00


/pci@0/pci@0/pci@1/pci@0/pci@1/pci@0/usb@0,2/storage@2
Unit 0 Removable Read Only device TSSTcorpCD/DVDW TS-T632ASR03


The Ultra5 does not have USB but we also have a SunFire V100 with two USBs in the back and Solaris 2.8 does support external USB drives although it is not clear if it has drivers. I found an inexpensive Plugable USB-SAS enclosure. I have always had good luck with Plugable devices so perhaps I'll get one.

I expected that the Solaris installation would simply format the drive as needed as part of the installation, similar to what Linux, Aix, Windows, etc... do. If none of this works, I'll risk upgrading the v100 from 2.8 to 2.10. I still will not have remote start/stop capability but I'll have a faster 2.10.

Thanks for your response!
Regards,
Paul

[invalid]

try

# format

and see what that shows

> I expected that the Solaris installation would simply format the drive as
>needed as part of the installation, similar to what Linux, Aix, Windows,
>etc... do.

I think it does but if you had something else installed on there it's not
uncommon for Solaris to ignore it. Unlike the rest and especially Windows
which happily tramples your drives, even ones you're not installing on.

[pa...@trifox.com]

Using the shell that the installation drops you into when it fails to find any disks,

# format
Searching for disks...done
No disks found!

#

Not surprising.

I went to a local liquidator who had a T5120 for sale and hooked up a laptop to ILOM. When I booted it to {ok} and typed

show-disks

it also did not show anything for the two disks that were installed. probe-scsi-all did not do anything either. I could not install a fresh Solaris on it because it complained about some out of date firmware.
Thanks,
Paul
Regards,
Paul

[John D Groenveld]

In article <3141f78c-1f0f-4044...@googlegroups.com>,

<pa...@trifox.com> wrote:
>I went to a local liquidator who had a T5120 for sale and hooked up a
>laptop to ILOM. When I booted it to {ok} and typed
>
>show-disks
>
>it also did not show anything for the two disks that were installed.
>probe-scsi-all did not do anything either. I could not install a fresh
>Solaris on it because it complained about some out of date firmware.

What version OpenBoot is installed?
From OK prompt, .version

What does show-devs reveal about your installed HBA?
ok reset-all
ok show-devs

John
groe...@acm.org

[pa...@trifox.com]

Hi John,
.version:
Release 4.30.4 created 2009/08/19 07:23

show-devs:
/ebus@c0
/pci-performance-counters@0
/pci@0
/niu@80
/cpu@3f
/cpu@3e
/cpu@3d
/cpu@3c
/cpu@3b
/cpu@3a
/cpu@39
/cpu@38
/cpu@37
/cpu@36
/cpu@35
/cpu@34
/cpu@33
/cpu@32
/cpu@31
/cpu@30
/cpu@2f
/cpu@2e
/cpu@2d
/cpu@2c
/cpu@2b
/cpu@2a
/cpu@29
/cpu@28
/cpu@27
/cpu@26
/cpu@25
/cpu@24
/cpu@23
/cpu@22
/cpu@21
/cpu@20
/cpu@1f
/cpu@1e
/cpu@1d
/cpu@1c
/cpu@1b
/cpu@1a
/cpu@19
/cpu@18
/cpu@17
/cpu@16
/cpu@15
/cpu@14
/cpu@13
/cpu@12
/cpu@11
/cpu@10
/cpu@f
/cpu@e
/cpu@d
/cpu@c
/cpu@b
/cpu@a
/cpu@9
/cpu@8
/cpu@7
/cpu@6
/cpu@5
/cpu@4
/cpu@3
/cpu@2
/cpu@1
/cpu@0
/virtual-devices@100
/virtual-memory
/memory@m0,8000000
/aliases
/options
/openprom
/chosen
/packages
/ebus@c0/serial@0,ca0000
/pci@0/pci@0
/pci@0/pci@0/pci@9
/pci@0/pci@0/pci@8
/pci@0/pci@0/pci@2
/pci@0/pci@0/pci@1
/pci@0/pci@0/pci@9/scsi@0
/pci@0/pci@0/pci@9/scsi@0/disk
/pci@0/pci@0/pci@9/scsi@0/tape
/pci@0/pci@0/pci@8/pci@0
/pci@0/pci@0/pci@8/pci@0/pci@a
/pci@0/pci@0/pci@8/pci@0/pci@9
/pci@0/pci@0/pci@8/pci@0/pci@8
/pci@0/pci@0/pci@8/pci@0/pci@2
/pci@0/pci@0/pci@8/pci@0/pci@1
/pci@0/pci@0/pci@8/pci@0/pci@9/SUNW,qlc@0,1
/pci@0/pci@0/pci@8/pci@0/pci@9/SUNW,qlc@0
/pci@0/pci@0/pci@8/pci@0/pci@9/SUNW,qlc@0,1/fp@0,0
/pci@0/pci@0/pci@8/pci@0/pci@9/SUNW,qlc@0,1/fp@0,0/disk
/pci@0/pci@0/pci@8/pci@0/pci@9/SUNW,qlc@0/fp@0,0
/pci@0/pci@0/pci@8/pci@0/pci@9/SUNW,qlc@0/fp@0,0/disk
/pci@0/pci@0/pci@2/scsi@0
/pci@0/pci@0/pci@2/scsi@0/disk
/pci@0/pci@0/pci@2/scsi@0/tape
/pci@0/pci@0/pci@1/pci@0
/pci@0/pci@0/pci@1/pci@0/pci@3
/pci@0/pci@0/pci@1/pci@0/pci@2
/pci@0/pci@0/pci@1/pci@0/pci@1
/pci@0/pci@0/pci@1/pci@0/pci@3/network@0,1
/pci@0/pci@0/pci@1/pci@0/pci@3/network@0
/pci@0/pci@0/pci@1/pci@0/pci@2/network@0,1
/pci@0/pci@0/pci@1/pci@0/pci@2/network@0
/pci@0/pci@0/pci@1/pci@0/pci@1/pci@0
/pci@0/pci@0/pci@1/pci@0/pci@1/pci@0/usb@0,2
/pci@0/pci@0/pci@1/pci@0/pci@1/pci@0/usb@0,1
/pci@0/pci@0/pci@1/pci@0/pci@1/pci@0/usb@0
/pci@0/pci@0/pci@1/pci@0/pci@1/pci@0/usb@0,2/hub@4
/pci@0/pci@0/pci@1/pci@0/pci@1/pci@0/usb@0,2/storage@2
/pci@0/pci@0/pci@1/pci@0/pci@1/pci@0/usb@0,2/storage@2/disk
/virtual-devices@100/rtc@5
/virtual-devices@100/console@1
/virtual-devices@100/random-number-generator@e
/virtual-devices@100/ncp@6
/virtual-devices@100/n2cp@7
/virtual-devices@100/channel-devices@200
/virtual-devices@100/flashprom@0
/virtual-devices@100/channel-devices@200/virtual-channel-client@1
/virtual-devices@100/channel-devices@200/virtual-channel@0
/virtual-devices@100/channel-devices@200/virtual-channel-client@2
/virtual-devices@100/channel-devices@200/virtual-channel@3
/openprom/client-services
/packages/obp-tftp
/packages/kbd-translator
/packages/SUNW,asr
/packages/dropins
/packages/terminal-emulator
/packages/disk-label
/packages/deblocker
/packages/SUNW,builtin-drivers

The drive is currently in slot 1. I tried it in slot 0, no difference.
Thanks,
Paul

[pa...@trifox.com]

> The Ultra5 does not have USB but we also have a SunFire V100 with two USBs in the back and Solaris 2.8 does support external USB drives although it is not clear if it has drivers. I found an inexpensive Plugable USB-SAS enclosure. I have always had good luck with Plugable devices so perhaps I'll get one.

Oops, I realize that is a USB-SATA, not SAS enclosure. It seems that there are no such animals out there.
Paul

[DoN. Nichols]

On 2016-08-15, pa...@trifox.com <pa...@trifox.com> wrote:
> Hi DoN,

> It's a brand new, never used drive: SAS Seagate Savio 300Gb. IIUC you
> are saying that it must be formatted by Solaris before Solaris can be
> installed on it. Seems a real "chicken-egg" problem or perhaps a way for
> Oracle to require users to purchase drives through them. When the
> installation fails (no disk found), it drops into a shell where I tried
> format but of course, there are no drives.

That should give you access to other tools to do what you need
-- but I'm not sure what those tools might be, yet.

> probe-scsi-all returns:
>
> {0} ok probe-scsi-all
> /pci@0/pci@0/pci@9/scsi@0
> Waiting for AAC Controller to start: . . Started
>
> AAC Kernel Version: 15583
>
>
> /pci@0/pci@0/pci@8/pci@0/pci@9/SUNW,qlc@0,1
> QLogic QLE2462 Host Adapter Driver(SPARC): 1.24 11/15/06
> Firmware version 4.00.26
> Fibre Channel Link down
> Possible causes: No cable, incorrect connection mode or data rate
>
> /pci@0/pci@0/pci@8/pci@0/pci@9/SUNW,qlc@0
> QLogic QLE2462 Host Adapter Driver(SPARC): 1.24 11/15/06
> Firmware version 4.00.26
> Fibre Channel Link down
> Possible causes: No cable, incorrect connection mode or data rate
>
> /pci@0/pci@0/pci@2/scsi@0
>
> MPT Version 1.05, Firmware Version 1.22.00.00
>
>
> /pci@0/pci@0/pci@1/pci@0/pci@1/pci@0/usb@0,2/storage@2
> Unit 0 Removable Read Only device TSSTcorpCD/DVDW TS-T632ASR03

Hmm ... this looks a bit different from my experience with this
(my system is up and acting as a server, and my wife would be upset
if I took it down to run some tests. :-).

But -- it suggests something like what I have experienced with
an X4150 (Intel CPUs) which has a hardware RAID card in it, and the
drive bays on the system (eight bays) are connected to two ports on the
RAID card.

For the drives to be visible to the system, it is necessary to
interrupt the boot at just the right time with a "CTRL-A" (IIRC) and
configure the firmware of the RAID card (a bunch of menus for that) so
*it* recognizes the drives, and passes them to the system CPUs. Change
a drive for a new one, and you have to do this again -- no swap a drive
in a booted system, format it, and then (perhaps) swap back. If your
T5120 has such a RAID card, there must be some way to do it, though
SPARC systems have a different way to do things than the booting through
the BIOS on the Intel/AMD systems.

Interestingly enough -- the RAID card won't see SATA drives,
just SAS, and if I bypassed the card and plugged the cables into the
system board, I think that it would only see SATA drives. (But I don't
know whether this applies to the T5120 or my T5220). I did see a T5220
on eBay with 16 drive bays instead of the eight that mine has which was
listed with a RAID card.

Have you downloaded the manuals for the T5120? Here are the
ones which I downloaded:

======================================================================
820-2177-14.pdf 820-4418-11.pdf E23172-01.pdf
820-2178-14.pdf 820-5839-11.pdf 820-2179-14.pdf
820-6683-13.pdf 820-2180-12.pdf 821-1592-10.pdf
820-2181-15.pdf
======================================================================

The E23172-01.pdf covers (in part):

======================================================================
1. Installing the StorageTek SAS RAID Internal HBA Into the SPARC
Enterprise T5120 and T5220 Servers 1
======================================================================

and it may describe how to configure the drives in it, too.

O.K. In "to verify the current devices":

======================================================================
{0} ok show-disks
> a) /pci@0/pci@0/pci@2/scsi@0/disk
> b) /pci@0/pci@0/pci@1/pci@0/pci@1/pci@0/usb@0,2/storage@2/disk
> q) NO SELECTION
> Enter Selection, q to quit: q
{0} ok
======================================================================

It shows that you need different internal cables to use the RAID
card or to use the controller built into the system card.

> The Ultra5 does not have USB but we also have a SunFire V100 with two
> USBs in the back and Solaris 2.8 does support external USB drives
> although it is not clear if it has drivers. I found an inexpensive
> Plugable USB-SAS enclosure. I have always had good luck with Plugable
> devices so perhaps I'll get one.

And -- I have found a PCI USB card to work in an Ultra-60, so
perhaps also in an Ultra-5. (Or is Ultra-5 still S-Bus? It has been a
while since I last ran one of those. :-)

Anyway -- the last page of that manual says:

======================================================================
4. Power on the server.
5. Perform any additional card-specific tasks.
======================================================================

which means from some other manual, not this one.

There are too many different manuals for me to dig through right
now. But you have the numbers, so you can download them and search
through them.

> I expected that the Solaris installation would simply format the drive
> as needed as part of the installation, similar to what Linux, Aix,
> Windows, etc... do. If none of this works, I'll risk upgrading the v100
> from 2.8 to 2.10. I still will not have remote start/stop capability but
> I'll have a faster 2.10.

If the drive was not sold as one for a Sun, it is probably fdisk
formatted. So you will need to use Some flavor of Solaris to first
re-format it to be a SPARC capable one -- unless you have a RAID card,
and you need to find out how to configure the RAID card to talk to the
drives. The RAID card does the formatting as needed, from within its
menu -- at least on the X4150, which again is an Intel CPU.

But -- my T5220 can handle both SATA and SAS drives, mixed if
necessary -- without the RAID card. So check what you have in the
machine(s) (you've got two, IIRC). If you have the RAID card, you may
have to pull it and get the shorter cables for the system-board included
SAS/SATA controllers.

There are a bunch of others who have apparently already
answered, and they may have the rest of the answers needed for you.

[DoN. Nichols]

On 2016-08-15, pa...@trifox.com <pa...@trifox.com> wrote:

It will likely work anyway. I've used similar ones with a SAS
drive connected to a Mac Mini to do an initial formatting of a drive
which the Suns would not see at first.

[pa...@trifox.com]

Hi DoN,
It is interesting that you mention RAID. After the install fails and I get the shell, I looked around the RAM directories and found /tmp/install_log.syslog. Dumping that, I saw

Aug 16 08:53:49 aac: [ID 477943 kern.info] NOTICE: aac driver 2.02.04-1, found card: Sun STK RAID INT(pci0x9005.285.108e.286) at 0xe00000

Searching on this, I found some Sun documents about the StorageTek Raid hw. But these are references on how to add drives to a RAID on a system that is already running. It seems to me that perhaps even before I "boot cdrom", I need to somehow be able to format the drive using some /SYS command so that when I "boot cdrom", the STK RAID driver can see the drive. Or maybe there is a different iso disk that is booted first in order to format drives.

Does this sound like I'm on the right track? I have a couple of the documents that you referenced and I will hunt down the rest of them.

Thanks also for the pointer on opening the box and poking around. In fact, I never bought the second T5120 that I looked at because it did not boot with the included drives and it acted the same as my system, i.e. show-disks did not show anything.
Thanks,
Paul

[tomatosoup]

I'm not sure how the RAID controller in the T5x20-series work, but on
older SPARC machines like the V440, you could list/create/delete RAIDs
only from Solaris via 'raidctl'.
Try to boot Solaris again via DVD, and do 'raidctl -l' - let's see if
you get any ouput.

Andy

[pa...@trifox.com]

Progress! The clue was RAID. I popped open the box, disconnected the SAS cable from the RAID card and plugged it into the motherboard. Start /SYS, start /SP/console, probe-scsi-all and it found the drive. Boot Solaris DVD, got to the end, disk error (corrupt label or not formatted, something like that). At the shell prompt, format found the drive and I am now two hours into an estimated five hour formatting session. More details as they come in!!

I have only one drive so RAID is a bit silly but I will try it again because the cable from the RAID card to the disk cage is 1) very long so I have it looped and 2) the connector sticks up and I am concerned about breakage from having the cover sit on it. Do the standard short cables have a 90 degree connector on them?

A big "Thank You" to all of you for your suggestions and pointers that got me this far!
Regards,
Paul

[tomatosoup]

On 2016-08-16 21:08:03 +0000, pa...@trifox.com said:

> I am now two hours into an estimated five hour formatting session. More
> details as they come in!!

There's no need to 'format' the whole disk - you just need to write a
label via the format utility to it, and you can use it immediately.

[pa...@trifox.com]

At this point, it is verifying. Is there any danger in stopping it at this point? The formatting is done:

format> label
Ready to label disk, continue? y

format> format

Ready to format. Formatting cannot be interrupted
and takes 228 minutes (estimated). Continue? y
Beginning format. The current time is Tue Aug 16 12:08:10 2016

Formatting...
99% complete (00:00:58 remaining) done

Verifying media...
pass 0 - pattern = 0xc6dec6de
5336/3/1875

It's not a big deal to wait, I have other things to do. If the estimate it correct, I have another hour or so to wait.
Paul

[tomatosoup]

On 2016-08-16 21:40:40 +0000, pa...@trifox.com said:

> At this point, it is verifying. Is there any danger in stopping it at
> this point? The formatting is done:

No, you can interrupt, this will not harm anything.

>
> format> label
> Ready to label disk, continue? y
>
> format> format
>
> Ready to format. Formatting cannot be interrupted
> and takes 228 minutes (estimated). Continue? y
> Beginning format. The current time is Tue Aug 16 12:08:10 2016
>
> Formatting...
> 99% complete (00:00:58 remaining) done
>
> Verifying media...
> pass 0 - pattern = 0xc6dec6de
> 5336/3/1875

As you can see above, this is only a read-test (Verfying media). You
could do some other things here to erase the disk, but in your case,
it's not needed.

>
> It's not a big deal to wait, I have other things to do. If the estimate
> it correct, I have another hour or so to wait.
> Paul

If you're not in a hurry you can wait for it to finish. The important
thing to do is 'label'; as you will install Solaris now, even that is
not necessary, as the installer will do it for you.

Andy

[invalid]

Isn't it better not to use RAID ever with ZFS?

[DoN. Nichols]

On 2016-08-16, pa...@trifox.com <pa...@trifox.com> wrote:
> Hi DoN,

> It is interesting that you mention RAID. After the install fails and I
> get the shell, I looked around the RAM directories and found
> /tmp/install_log.syslog. Dumping that, I saw

> Aug 16 08:53:49 aac: [ID 477943 kern.info] NOTICE: aac driver
> 2.02.04-1, found card: Sun STK RAID INT(pci0x9005.285.108e.286) at
> 0xe00000

That is it, then. You need to find out how to do the
configuration (if nobody else has already posted it). On my X4150, of
course, the RAID card has an X86 program in its part of the BIOS ROM
(on the card), and the BIOS searches it out and offers it at the right
time.

Obviously, that code won't work on a SPARC, and the card needs
some FCODE in a ROM on the card -- and some magic way to invoke it.
(There are so many commands buried in OpenBoot I have no how to find
such a command.) But I do strongly suspect that there is some way to do
it.

> Searching on this, I found some Sun documents about the StorageTek
> Raid hw. But these are references on how to add drives to a RAID on a
> system that is already running. It seems to me that perhaps even before
> I "boot cdrom", I need to somehow be able to format the drive using some
> /SYS command so that when I "boot cdrom", the STK RAID driver can see
> the drive. Or maybe there is a different iso disk that is booted first
> in order to format drives.

That could be -- but I consider fcode in a ROM on the RAID card
as the most likely. Just not sure how to find it.

If Casper Dik is here, I'll bet that he knows. :-)

Note that I am just a hobby user who likes to pick up Sun
computers at hamfests and from eBay. I was a sysadm for networked Suns
the last few years at work -- but that was shortly before the UltraSPARC
systems came around. :-)

> Does this sound like I'm on the right track? I have a couple of the
> documents that you referenced and I will hunt down the rest of them.

Note that some of them simply refer to how to mount the computer
in the rack -- these are all the ones which I downloaded for the T5220
(and as I say, most of them apply to the T5120 as well.)

> Thanks also for the pointer on opening the box and poking around. In
> fact, I never bought the second T5120 that I looked at because it did
> not boot with the included drives and it acted the same as my system,
> i.e. show-disks did not show anything.

Note that one of the options in the BIOS menu for the RAID on
the X4150 is "secure erase", which of course also clears the operations
which the RAID card does to prepare the drives. (Even assuming that the
drives were in that system when it was in service. :-) A lot of people I
know would not trust the "secure erase" option -- and while I haven't
tried it on my system, I'll bet that it takes forever for a reasonably
large drive (e.g. 146GB and larger. :-)

[DoN. Nichols]

On 2016-08-16, tomatosoup <tomatosou...@lostcreations.org> wrote:
> I'm not sure how the RAID controller in the T5x20-series work, but on
> older SPARC machines like the V440, you could list/create/delete RAIDs
> only from Solaris via 'raidctl'.
> Try to boot Solaris again via DVD, and do 'raidctl -l' - let's see if
> you get any ouput.

That sounds good. I hope that it will work for him. I do find
"raidctl" on my SPARC Solaris 10 systems -- and will have to see what it
does on the X86 (X4150) system -- once I power it back up. (It is
powered down at the moment because of a bunch of thunderstorms and the
fact that there is not enough capacity in the UPS to support it along
with the other systems already in use. :-) At least it is present in a
system with X86 Solaris 10 which does not have a RAID controller, so
that seems to be the way to go.

Since you don't have the man page working booted from the DVD,
but IIRC, you have other Solaris 10 systems, go to one of those to read
(or print) the man page for raidctl.

So -- it seems that "tomatosoup" solved your initial problem.

[DoN. Nichols]

On 2016-08-17, invalid <add...@is.invalid> wrote:
> Isn't it better not to use RAID ever with ZFS?

Well ... I *had* to use RAID with a StorEdge T3+ while I was
using it -- but I had it configured to provide one large partition with
a hot spare disk, and then used ZFS to partition it as I wanted to get
most of the benefits of ZFS.

But, yes, I prefer ZFS over hardware RAID.

And I'm not sure how easy it would be for the OP to convert his
system to non RAID. It only involves pulling the card, and running new
cables from the drive bays to the controller on the system board -- but
those require a different length of specialized cable, which he might not
have access to -- and tucking the longer cables into the necessary loops
to reach the needed system board controllers might block the airflow
somewhat.

I've got the RAID controller on the X4150 configured to supply
all the drives as individual drives, so ZFS works as designed for the
moment. On that system (based on what I have read so far), apparently
the RAID card is needed to talk to SAS drives, and the absence of the
RAID card is needed to talk to SATA drives.

Enjoy,

[invalid]

On 2016-08-18, DoN. Nichols <BPdnic...@d-and-d.com> wrote:
> On 2016-08-17, invalid <add...@is.invalid> wrote:
>> Isn't it better not to use RAID ever with ZFS?
>
> Well ... I *had* to use RAID with a StorEdge T3+ while I was
> using it -- but I had it configured to provide one large partition with
> a hot spare disk, and then used ZFS to partition it as I wanted to get
> most of the benefits of ZFS.
>
> But, yes, I prefer ZFS over hardware RAID.
>
> And I'm not sure how easy it would be for the OP to convert his
> system to non RAID. It only involves pulling the card, and running new
> cables from the drive bays to the controller on the system board -- but
> those require a different length of specialized cable, which he might not
> have access to -- and tucking the longer cables into the necessary loops
> to reach the needed system board controllers might block the airflow
> somewhat.

Seems to me it might be faster to pull the RAID card and go plain JBOD if at
all possible. Even if he gets the RAID going it's going to interfere with
ZFS. ZFS is the best reason to run Solaris on Sun hardware!

> I've got the RAID controller on the X4150 configured to supply
> all the drives as individual drives, so ZFS works as designed for the
> moment. On that system (based on what I have read so far), apparently
> the RAID card is needed to talk to SAS drives, and the absence of the
> RAID card is needed to talk to SATA drives.

Ok, I really don't know. I thought there were times when even a passthrough
RAID setup either caused actual problems or reduced recoverability or
performance. I am no expert on this. Hopefully people who know from
experience can fill us in.

[pa...@trifox.com]

I formatted ZFS with the original long RAID cable plugged into the MB controller socket. The cable is looped and I think the cover is resting on the wires coming out of the plug. As for the airflow, the system is not run 24/7 so that should not be a big issue. Does anyone know the part number of the shorter cable? I hunted around but never found a part number or anything equivalent. Does it have a 90 degree plug so that the wires does not stick up so high? I have only one drive so RAID would be a little silly and the stuff that I have on the T5120 is easily reloadable so it's not a big deal.

Thanks again to everyone who responded!
Paul

[John D Groenveld]

In article <a3a87762-246a-48b7...@googlegroups.com>,

<pa...@trifox.com> wrote:
>I formatted ZFS with the original long RAID cable plugged into the MB
>controller socket. The cable is looped and I think the cover is resting
>on the wires coming out of the plug. As for the airflow, the system is

Sun's old manual describes how the cable should be routed from
mainboard's SAS controller to the disk backplane:
<URL:https://docs.oracle.com/cd/E19839-01/820-2181-15/820-2181-15.pdf>

>not run 24/7 so that should not be a big issue. Does anyone know the
>part number of the shorter cable? I hunted around but never found a part

Possibly
SECX9SA1Z [S] 560-2920 [F] Mini SAS Cable Kit

John
groe...@acm.org

[John D Groenveld]

In article <np3tkk$1uas$1...@gioia.aioe.org>, invalid <add...@is.invalid> wrote:
>Ok, I really don't know. I thought there were times when even a passthrough
>RAID setup either caused actual problems or reduced recoverability or
>performance. I am no expert on this. Hopefully people who know from
>experience can fill us in.

Yes, the Adaptec RAID controller is slow, even configured
with single disk RAID-0 volumes for ZFS use.
Don would see better performance by swapping in a LSI SAS HBA.

And recovery is simplified by removing the Adaptec firmware step
when replacing failed drives.

John
groe...@acm.org

[DoN. Nichols]

On 2016-08-18, pa...@trifox.com <pa...@trifox.com> wrote:

> I formatted ZFS with the original long RAID cable plugged into the MB
> controller socket. The cable is looped and I think the cover is resting
> on the wires coming out of the plug. As for the airflow, the system is
> not run 24/7 so that should not be a big issue. Does anyone know the
> part number of the shorter cable?

Both the longer and shorter cables are identified in the:

E23172-01.pdf

document. Is yours a 4-disk or an 8-disk system? There is one cable
for 4-disk and two for 8-disk.

Nope! Only the cables to be installed are identified, on a HBA
by HBA basis -- ot those which are to be removed.

4-drive
530-4119 CN0
8-drive
530-4118 CN0
530-3893 CN1
8-drive
530-4119 CN0
530-3892 CN1
16-drive
530-4119 CN0

No info as to what the numbers of the cable(s) to be removed
might be. Sorry.

I could not shut down the T5220 which is acting as server
without my wife being quite upset. Sorry.

> I hunted around but never found a part
> number or anything equivalent. Does it have a 90 degree plug so that the
> wires does not stick up so high? I have only one drive so RAID would be
> a little silly and the stuff that I have on the T5120 is easily
> reloadable so it's not a big deal.

A bit more of a problem, with raidctl, as reading the man page
shows a minimum of two drives in a RAID-0 or a RAID-1 configuration. No
mention of how to pass through individual drives as I have on the X4150.

> Thanks again to everyone who responded!
> Paul

Good Luck,