IRIX and Software

Hello,

long story short, I got in possession of another SGI Fuel (whose psu is not giving problems up to now, crossing fingers).
It has still the original installation from the company it used to belong (minus their licenses, ach... there's some catia 4.1 installation whose license was apparently removed - unless it expired, obviously ).
It's a sort of basic, I guess, setup :
R14k 600mhz
1.5gb ram ( strangely 1 module is premium.. would have expected identical pairs )
18gb hd
v10

One monitoring ic went faulty during the 2 - 3 years the previous owner kept it stored and disconnected from any power line ( realized is common to happen on older batches ) so had to disable env monitoring.

The os is a 6.5.15m with 2 or 3 patches installed.

I'm wondering what would be the best way to proceed...
keeping it at 6.5.15m , or cloning the disk and upgrading it... but if going this way I wonder what would be the best way to avoid issues (and breaking some sw already on it).
Like :
installing missing things now , then upgrade ?
wait to install the missing after upgrading ?
upgrade straight from .15 to .30 ? or is better to move up to an intermediate and only then move to .30 ?

Any things I may have forgotten to ask/mention ?

It depends on whether the disk contains interesting data for the sake of preservation. If not, upgrade to 6.5.30m or wipe and install fresh.
For an upgrade from 6.5.15m to 6.5.30m you need patch 5086 installed first: http://www.vdheijden-messerli.net/sgist ... 86.tardist

BTW, the sensible choice of IRIX version on a Fuel would be 6.5.30m, but a minimum of 6.5.22m is required to run nekoware. This is because that version is the ultimate version for older SGI systems and a sizeable part of nekoware is built against that version.

The media can be located on Ebay, or downloaded from Archive.org if you don't mind some CD image juggling and setting up DINA or a Linux machine as install server.

Check out Ian's site at http://www.sgidepot.co.uk/sgi.html for extra stuff and information.

ok thanks.
I guess it depends also what we can define interesting data

Search very hard for that Catia licence chap, check out what's in all the trash, those things are like gold.

CATIA uses IBM LUM licensing which has it's roots in NetLS. On IRIX a nodelock license should be in /var/lum or /opt/lum/ls/os/svr4.sgi/bin.

If this system used a floating license you'll never find it, of course because you don't have the floating license server.

jan-jaap wrote: CATIA uses IBM LUM licensing which has it's roots in NetLS. On IRIX a nodelock license should be in /var/lum or /opt/lum/ls/os/svr4.sgi/bin.

If this system used a floating license you'll never find it, of course because you don't have the floating license server.

Very true, but I do tend to look in peoples trash, you never know what goodies await you.

Personally, I'd always recommend cloning the original disk first. Just always gives you something to go back to if you need it, then gives you the freedom to play around with the Fuel - knowing you've got a copy of all the original data in case you do want to find an old license or something. Of course, I've got about 25 spare hard drives kicking around, so I've no shortage of drives for this

thanks for the suggestions.
I'm planning to slowly check the nooks and crannies of the filesystem to see if there's something particular. The previous owner, to my knowledge :
-got it
-looked at what was inside enough to take pics
-disassembled
-took pics of the parts
-reassembled
-put away

I'm fairly sure any action from his side may have been limited to root password reset
any utility for irix to recover deleted files, in case ? ( not just trashed )

jan-jaap wrote: CATIA uses IBM LUM licensing which has it's roots in NetLS. On IRIX a nodelock license should be in /var/lum or /opt/lum/ls/os/svr4.sgi/bin.

If this system used a floating license you'll never find it, of course because you don't have the floating license server.

update

sorry for taking time, but had no time ( pun not intentional )
under /var/lum
there's indeed a bit of things, including a
nodelock-sgi115.txt with a catia license inside for
MECH
ANS (KIN)
SDS (ASH)
expiration date : 2002 / 2003

late night update

ProductVision MockUp present , EAI3.0

Traces that this unit belonged to Bertone engineering, the car designing company

maybe I should make a new thread but...

I just cloned the boot disk using betxen script, changed the id of the cloned to 1 and trying to boot from it I get a

unable to execute dsck(0,1,8)/sash , no such device
unable to load bootfile

printevn from prom gives :
systempartition=dsck(0,1,8)
osloadpartition=dsck(0,1,0)
osloader=sash
osfilename=unix


suggestions ? means the source hd has 2 partitions ?

All IRIX disks are required to have multiple partitions. The partition numbered 8 is the disklabel and contains the boot loader program (sash). Without it, you cannot boot IRIX from the disk.

The devices used should be
They are case and spelling sensitive. What you wrote won't work.

they were written as you wrote, just I was too lazy to do an 1:1 copy here.
Mainly was trying to understand what I missed / was missed by the scrip... assumed the script would have done all the required steps to clone the original disk.

And what I should do to fix it ( consider this is the first time I try cloning.. )

These

dvhtool -v get sash sash /dev/rdsk/dks${orig_cntrl}d${orig_drv}vh
dvhtool -v get ide ide /dev/rdsk/dks${orig_cntrl}d${orig_drv}vh
dvhtool -v creat sash sash /dev/rdsk/dks${target_cntrl}d${target_drv}vh
dvhtool -v creat ide ide /dev/rdsk/dks${target_cntrl}d${target_drv}vh

were in the script, and were executed. ide was not present in the original disk, but sash didn't give an error
and I added

dvhtool -v get symmon symmon /dev/rdsk/dks${orig_cntrl}d${orig_drv}vh
dvhtool -v creat symmon symmon /dev/rdsk/dks${target_cntrl}d${target_drv}vh

that was found and copied

---

ehm.. seems I have an issue.

I simply plugged back the original disk, and is not booting ... now I get the same error about sash not found

---

progress :

esf read error, bad count
esf read error, bad count
no file system found for "unix".
unable to execute dksc(0,1,0)unix: media not loaded

autoboot failed
dksc(0,1,0)unix: media not loaded

latest edit : the abovementioned progress is relative to the cloned hd. The original one is ok.

That script doesn't really clone disks. It doesn't set up the disklabel on the target disk, you are required to do that yourself.
If you didn't partition the target disk before running the script nothing will work.

mmm I did initialize the disk using the graphical disk manager (aka there's a label)... but now I'm not sure if I did partition it... assumed it was similar to dd, fsdump.

oh well... live and learn.

btw, the original hd has a script in the root, made by the company it was being used for, to clone ( and for multiple installations )
The script is a lot more simple and with less steps than the ones I saw here

xfs_copy -d /dev/dsk/dks0d1s0 /dev/dsk/dks0d2s0

and that's all ( aside being followed by 2 entries to edit the hosts and sys_id ).
I have also the feeling that they cloned a disk originally installed in a server, since strangely there's no ide on it, so no diag tools to launch from prom

ide doesn't need to be in the volume header, since it can be loaded from the filesystem with sash. It's used rarely enough that the additional steps aren't a problem (need to enter the command prompt and type "boot usr/stand/ide" instead of just clicking "diagnostics")

sorry to say, but

/usr/stand

doesn't exist on the disk ( I dumped the list of everything )
The only stand present is in the root
/stand

with fx and sash inside ( and now with symmon, too )

mazzinia wrote: mmm I did initialize the disk using the graphical disk manager (aka there's a label)... but now I'm not sure if I did partition it

for low level stuff your best bet always is the terminal. here's how you prepare a disk as system disk:

then format it:

up next: the volume header


replace the s0d3 parts with whatever you need. the s0 is the controller number and the d3 is the disk id.
now mount the disk's s0 partition (dks0d3s0 in the example here) and start dragging the data over. xfsdump/restore is the recommended way but tar, cpio and alike should work as well.
after that unmount the disk and you're good to go

in theory those 3 steps were done

gui to initialize and put a label
script formatted and copied the header ( sash and symmon, in my case )

I checked the headers of the 2, and were identical

the only step I think I forgot to do, was to check the source partitioning, and to repartition the target

mmmm, indeed the 2 partitions are not identical

sgi115 2# prtvtoc
Printing label for root disk

* /dev/root (bootfile "/unix")
* 512 bytes/sector
* Unallocated space:
* Start Size
* 35546876 296794
*
Partition Type Fs Start: sec Size: sec Mount Directory
0 xfs yes 8392704 27154172
1 raw 4096 8388608
8 volhdr 0 4096
10 volume 0 35843670

sgi115 3# prtvtoc /dev/rdsk/dks0d9vh
* /dev/rdsk/dks0d9vh (bootfile "/unix")
* 512 bytes/sector
Partition Type Fs Start: sec Size: sec Mount Directory
7 xfs yes 4096 71128863
8 volhdr 0 4096
10 volume 0 71132959

sgi115 4#