[robespierre]

I think you're going down the wrong track trying to fix a directory problem with fx, which is a much lower-level tool.
The misplaced directory can only be found by mounting the volume in a working operating system (or the miniroot). fx cannot be used to find or move it. Just forget it.

[jwhat]

Hi jmjsgi & robspierre,

I was not suggesting to fix file sytems using fx, just use fx to confirm that the file system exists on disk (which based on info provided it should)

Until jmjsgi mounts the file systems and has a look at it (which is why he asked about mounting read only), then we are just guessing.

Simple steps:

1. Understand the disk layout (fx & labels)
2. Mount the volumes (root & optionally usr)
3. Poke around :-) ....

Nice thing about Unix based OSes is that you can rebuilt a system from file system "bits" using a combination of new installation and restoring the "bits" that are available and recoverable. I have done this with both IRIX and Ubuntu where file system/disks got partial contents "removed".

ie no Windows Registry anti-pattern ;-) , where things work via hidden information.

jmjsqi - suggest you ssh or telnet into the machine you have installed the disk in and then you can cut/paste terminal contents/logs rather than posting "pictures", it will make diagnosis easier.

Cheers from Oz,


jwhat/John.

[jmjsgi]

thank you very much for your feedback.
Thanks to jwhat for the command lines.
I checked the partition with the "part" command in "fx".



Here is the information on my disk:

   

https://imgur.com/wzLPyNZ

-----partitions------
part      type                         blocks                                                             megabytes           (base+size)
  0:        xfs                    266240 + 286483248                                             130 + 139884
  1:        raw                       4096 + 262144                                                       2 + 128
  8:        volhdr                        0 + 4096                                                            0 + 2   
10:        volume                       0 + 286749488                                                  0 + 140014

I created the BOOT directory in tmp and used the command to mount the disk in the BOOT directory.
"mount -t xfs -o ro /dev/dsk/dksXdYs0 /tmp/BOOT"
the system did not want to mount the disk as read-only.

   

https://imgur.com/6G6YBAY

So I typed the command:
mount -t xfs -o rw /dev/dsk/dksXdYs0 /tmp/BOOT

my disk is mounted correctly

I was able to access the contents of my disk (tezro) and comparing the disk of my Onyx, I saw that the var directory had disappeared.
I found it in the "user_data" directory.
I moved it to the root of my mounted tezro disk (tmp/BOOT).
Then I wanted to unmount the disk with umount:
umount -t xfs -o rw /dev/dsk/dksXdYs0 /tmp/BOOT
I got an error message (see photo)

   

https://imgur.com/I0h03AH

umount -a -t xfs /dev/dsk/dks0d2s0/ /tmp/BOOT
or
umount -a -t xfs /dev/dsk/dks0d2s0/
I couldn't take it apart.

   

https://imgur.com/bvkZ3Bn

Can you give me the command so I can do this without affecting the dks0d2s0 disk and even less the one on my onyx.

Thank you.

[jwhat]

Hi jmjsgi,

to unmount you use the directory that mount is on not the device.

So if mount is on /tmp/BOOT then:

umount /tmp/BOOT

Sounds like you have found the problem, but it is very strange.

I see you have put the mount in /etc/fstab , given that this is just temporary mount for diag/fix, I would not bother with that ;-)

Do a check on your disk for "wayward" symbolic links on in your top directory "/" by doing an "ls -la" and looking to see if there are any "strange" links.

This is sometimes the case when you "inherit" a machine and the prior owner has moved thiings about and put symbolic links to make sure it will works.

On my Fuel it had lots of links in the various application install directories to make it stuff everything into a single place.

This took quite a bit of work to unwind.

Does the machine that the boot disk come from have mutliple disks installed ?

It might be that one of these is used for /var mount point.

On the boot disk, check the fstab to see what is being mounted:

cat /tmp/BOOT/etc/fstab

I see the read only mount failed as disk was not clean, this typically happens then you just power off machine (ie pull the plug), without going through proper shutdown (either via button or shutdown command).

IRIX does auto fschk / xfs_repair (file sytem check) on boot, but it obviously must have write access to fix things.

Out of curiosity does read only mount work now it has been cleaned up ?
I should but empirical evidence provides the definitive answer ;-)

Cheers from Oz,


jwhat/John

[jmjsgi]

hi jwhat,
I just checked the contents of my disk, disassembled it and installed it on my tezro.
I'm super happy because the station starts perfectly, softimage and flame too and what's more, before it took forever to stop whereas now, it stops in less than 10 seconds.

Many thanks for your help.
If you need anything in France or nearby, don't hesitate to ask.

Jmjsgi.