help! can't excute

UNIX for Advanced & Expert Users
1

I can't excute any commands in /usr/bin

when I try to excute commands in /usr/bin
the error message is displayed " No Such device or
address"

Can you Help?

Thanks in Advance

2

Can execute them using the absolute address?

i.e. /usr/bin/<command>

Is /usr/bin in your list of paths?

echo $PATH

If you find it's not in your paths, check your .profile

You could also check the permissions of the commands in /usr/bin

If it's not any of these, post back with a bit more detail if possible

3

I have already checked the above suggestions but none of commands in the /usr/bin directory can be executed, no changes has been made on server.

4

I believe in other posts you have mentioned you run Solaris.

Was /usr a separate file partition?

If it was, you may have a problem with it mounting. Check to see if it's mounted with the df command (that should be out of /bin/df).

If it was not separate, then suggest a boot -s from cdrom...you can bring your server down with sync, sync, sync, halt and it probably won't hurt it as bad as just a poweroff since init 0 will run through the rc kill scripts and possibly hang (due to the problem).

Was the server up and running when this happened? Can you get to the /var/adm/messages to cat the file? You can also try dmesg which may show any error. Or if you were using a volume manager (SDS or Veritas) those commands may show something.

5

RTM has a good idea actually.

I remember right after i did an install the disk had gone bad but it still looked mounted but i could not get any of the data but i could see it all. it was quite weird.

try unmounting that FS and then remounting. If you get errors remounting it (and its not from your command error) then sounds like you might have a currupt FS.

6

If this is SunOS, can you verify that "/usr/bin/uname -a" fails and "/sbin/uname -a" succeeds? The former is dynamically linked while the latter is statically linked. If the dynamically linked version is failing you have some kind of shared library problem. By running truss on the dynamically linked version, I notice the open on /dev/zero which looks like the only opportunity that I see to generate the particular error that you are reporting. Can you verify that /dev/zero is present? What shell are you using?

7

Thanks everyone.

I rebooted the server and the server had problem coming back up. it seems server cannot see the disk

from OK prompt I issued probe-scsi-all

and the this displayed the error message "Fast Data Access MMU"

I followed the procedure on sunsolve and this solved the problem

http://sunsolve.sun.com/pub-cgi/retrieve.pl?doc=finfodoc%2F28291&zone_32=fast%20data%20access%20mmu

Note
I followed the procedure without reinstalling the OS and this
sovled the problem

Thanks again