Nekochan Net

Someone asked me to have a look at some old mageto-optical disks containing lot of backup scientific data they don't want to lose, because it does not work anymore on their workstation.
The drive, a Maxoptix Tahiti IIm, was previously connected to a R5K180 O2 running Irix 6.3 and mounted that way:
The system always says ther are I/O errors and can't read the disks. The cartridges are Maxoptix 1Gb 512bytes/sector erasable ones.

I brought the drive home. On my machine (R5K180 O2 runnung Irix 6.5), I tried with the same options, then simply without success. When I insert the disk, the "optical" icon on the desktop chandes (now with a disk half inserted), but I can't access it when double-clicking (see picture): says "magneto-optical disk type still being identified". I also tried to mount it from the CLI. Not better.

Unfortunately I didn't found how to make it work, neither on the forum nor on the web. Does one of you have experience with these MOD under Irix ? Diego ?

Thanks in advance for your help.

_________________
4xR12k 400, 8Gb, IR2E, 2x18+3x73GB HD
16xItanium2 1.6, L2 9MB
R16k 800 V12, 2Gb, M-Audio, 36+147GB HD, 3Dconnexion SpaceMouse Classic
Dual R14k 600 V6, 2Go, HD (173Go, 34Go)
R10k 400 MXE, 1280Mo / 2xR10k 300 SSE,...
R10k 195, 512Mo / R5k 180, 512Mo / R5k 180->200 motherboard and PM only
R10k 195, HighImpact, 160Mo/ R4400 125, Extreme, 160Mo
twin O200, 4xR12k 270, 2Go, HD (4x18Go)
4D70GT... my very first one (now property of musée bolo and the foundation mémoires informatiques)

Don't know how helpful it might be to your situation, but there's a 1995 discussion in the comp.sys.sgi.hardware archives on problems mounting a Maxoptix Tahiti 3 (under IRIX 5.3). Not especially topical, but it does included advice from the heavy hitters of the comp.sys.sgi era - Dave Olsen and Alexis Cousein. The thread includes a mention that the Maxoptix Tahiti 2 and 3 don't properly identify themselves as MO drives, and some suggestions to get them working/recognized.
http://groups.google.com/group/comp.sys ... 8451fe8db7

and this even older IRIX Magneto-Optical FAQ suggests using to manually mount the drive.
http://giga.cps.unizar.es/~spd/pub/irix ... ptical-faq

On the odd chance the dip-switches got bumped out of place somewhere along the line (and the even odder chance the config was the same as used with Sun), there's mention of the switch positions used when a Maxoptix Tahiti 2m is connected to Sun hardware here: http://www.sunmanagers.org/archives/1993/1683.html

_________________
***********************************************************************
Welcome to ARMLand - 0/0x0d00
running...(sherwood-root 0607201829)
* InfiniteReality/Reality Software, IRIX 6.5 Release *
***********************************************************************

Thanks recondas.

I'd seen that discussion, but hadn't tried as it was for 5.2/5.3. Finally did yesterday what is suggested ("/etc/umount" and jumper change) but without success.
At least, the drive is detected in the PROM and by hinv. When a disk is inserted, it is also detected. Then, I tried to disable mediad on this device... no change. I haven't found any info whether 6.5's GUI is functional or if it is the same as for 5.x.

What is strange is that there never is /dev/dsk/dks1d5s7, only /dev/dsk/dks1d5vol. I also tried to mount it as hfs using the whole raw device ("/etc/mount -t hfs /dev/rdsk/dks1d5vol /TAHITI")...

Don't know if I can try anything else or if I should look for another drive.

_________________
4xR12k 400, 8Gb, IR2E, 2x18+3x73GB HD
16xItanium2 1.6, L2 9MB
R16k 800 V12, 2Gb, M-Audio, 36+147GB HD, 3Dconnexion SpaceMouse Classic
Dual R14k 600 V6, 2Go, HD (173Go, 34Go)
R10k 400 MXE, 1280Mo / 2xR10k 300 SSE,...
R10k 195, 512Mo / R5k 180, 512Mo / R5k 180->200 motherboard and PM only
R10k 195, HighImpact, 160Mo/ R4400 125, Extreme, 160Mo
twin O200, 4xR12k 270, 2Go, HD (4x18Go)
4D70GT... my very first one (now property of musée bolo and the foundation mémoires informatiques)

Maybe try to 'dd' the device to a harddisk? If you get I/O errors you know the disk or the drive (or both) have issues. Optical media don't exactly last forever, and this is early 90's tech.

_________________
Now this is a deep dark secret, so everybody keep it quiet
It turns out that when reset, the WD33C93 defaults to a SCSI ID of 0, and it was simpler to leave it that way... -- Dave Olson, in comp.sys.sgi
Currently in commercial service: (2x)
In the museum: almost every MIPS/IRIX system.
Wanted: GM1 board for Professional Series GT graphics (030-0076-003, 030-0076-004)

I'll try that.

Do you think it is better to use if=/dev/dsk/dks1d5vol or one of the /dev/rdsk device ? What is the difference ? Both of them are present- At first I thought the "r" meant removable (no, please don't laugh)... but it seems to mean "raw device", isn't it ?

What is the impact of the "bs" (block size) option ?

_________________
4xR12k 400, 8Gb, IR2E, 2x18+3x73GB HD
16xItanium2 1.6, L2 9MB
R16k 800 V12, 2Gb, M-Audio, 36+147GB HD, 3Dconnexion SpaceMouse Classic
Dual R14k 600 V6, 2Go, HD (173Go, 34Go)
R10k 400 MXE, 1280Mo / 2xR10k 300 SSE,...
R10k 195, 512Mo / R5k 180, 512Mo / R5k 180->200 motherboard and PM only
R10k 195, HighImpact, 160Mo/ R4400 125, Extreme, 160Mo
twin O200, 4xR12k 270, 2Go, HD (4x18Go)
4D70GT... my very first one (now property of musée bolo and the foundation mémoires informatiques)

you generally should use the raw device when copying disks with dd, it's faster and safer.
the blocksize options control dd's blocking factor; the size of the chunk it attempts to read and/or write. (partial chunks are handled too, just slightly differently.) The difference is that controls the size of the physical read() ONLY on raw devices. The "regular" disk device sets its own read size. You can set the bs to the sector size of the media, I think that would be best for error recovery. so bs=512 or 1024, etc. This will be slower since more read commands will be issued separately. For the best speed use bs=1024k or more for fast devices.

_________________

Thanks for the explanation. I'll then use "dd bs=512 if=/dev/rdsk/dks1d5vol of=/MO_disks/disk[X]_face[A/B].image" to put all chances on my side.

Is it possible to mount this image later on, as it is the complete volume, not only the 7th partition ?

_________________
4xR12k 400, 8Gb, IR2E, 2x18+3x73GB HD
16xItanium2 1.6, L2 9MB
R16k 800 V12, 2Gb, M-Audio, 36+147GB HD, 3Dconnexion SpaceMouse Classic
Dual R14k 600 V6, 2Go, HD (173Go, 34Go)
R10k 400 MXE, 1280Mo / 2xR10k 300 SSE,...
R10k 195, 512Mo / R5k 180, 512Mo / R5k 180->200 motherboard and PM only
R10k 195, HighImpact, 160Mo/ R4400 125, Extreme, 160Mo
twin O200, 4xR12k 270, 2Go, HD (4x18Go)
4D70GT... my very first one (now property of musée bolo and the foundation mémoires informatiques)

IRIX doesn't support loopback mounting, but you can always dd the image to an unused SCSI disk and mount that.

A Linux system configured with support for SGI disklabels and EFS filesystem can mount EFS disk images. I keep image copies of all my IRIX CDROMs on my fileserver (Debian 'stable')

_________________
Now this is a deep dark secret, so everybody keep it quiet
It turns out that when reset, the WD33C93 defaults to a SCSI ID of 0, and it was simpler to leave it that way... -- Dave Olson, in comp.sys.sgi
Currently in commercial service: (2x)
In the museum: almost every MIPS/IRIX system.
Wanted: GM1 board for Professional Series GT graphics (030-0076-003, 030-0076-004)

I think the media lasts forever but the drives were dying even before they left the factory

I had one of those Maxoptic Tahitis, never could get it to work. You might try on some NeXT groups. The M-O drive was a fixation with Mr Jobs early on, the NeXT guys have a lot of experience with getting data off the damned things.

conv=noerror or ignerror might be good for flaky media; you might get SOME of the data out of it even if the disk is occasionally unreadable

_________________
halo, octane
N.B.: I tend to talk out of my ass. Do not take it too seriously.

There's also ddrescue in /beta. If multiple attempts of reading bad sectors succeeds sometimes, that may work better than dd conv=noerror. I've packaged it, but haven't really needed it in the end.

Thanks for all these sugestions.
I've done a lot of tests this week-end without more success.

dd does not work and always gives: whichever disk and side I try (3 disks at home, written between '94 and '98).

ddrescue does not work either. The machine is not connected on the network, so I wont copy the output here, but the summary is the same as for dd...

Whatever command I use (dd/ddrescue), the console says:

Seems strange that every disk is corrupted. MO is said to last for 50 years, so let's reasonably say 25. One of them should still work at least.

Do you have other ideas ?

_________________
4xR12k 400, 8Gb, IR2E, 2x18+3x73GB HD
16xItanium2 1.6, L2 9MB
R16k 800 V12, 2Gb, M-Audio, 36+147GB HD, 3Dconnexion SpaceMouse Classic
Dual R14k 600 V6, 2Go, HD (173Go, 34Go)
R10k 400 MXE, 1280Mo / 2xR10k 300 SSE,...
R10k 195, 512Mo / R5k 180, 512Mo / R5k 180->200 motherboard and PM only
R10k 195, HighImpact, 160Mo/ R4400 125, Extreme, 160Mo
twin O200, 4xR12k 270, 2Go, HD (4x18Go)
4D70GT... my very first one (now property of musée bolo and the foundation mémoires informatiques)

These disks used to be formatted with different blocksizes. IRIX (or even the drive) may be having a problem if, for example, the disks have 256, 1024, or 2048 byte blocks.

_________________

From what I remember MO discs are hard sectored... the sector patterns are molded into the polycarbonate.
The OP said his discs are 512 byte.

The drive could also be bad. Or the discs could have "laser rot"...

_________________

My money is on "drive is bad." I've never seen a Maxoptix that worked. Spent a lot of time trying, too. After a while Weird Stuff would see me coming and jack up the prices on all their old broken Tahitis.

I also sort of remember 2048 being block size on those ... but Al Zeimer was my grandpa so ....