Lost connnection to the server

[Harry Yu]

Fresh problem. I lost connection to our HP-UX server just couple minutes ago. I could not log in as a root. From a client system, when I tried to connect via Telnet, it disconnects me right away. On the server's terminal, I entered root and password and the server would reject it. None of the user accounts work. Users that are logged in are still able to work. However, once they logged out, they can't log back in. I was adding a new user in SAM before this happen. I did not delete any account from SAM. I don't know what is going on and just not able to do anything. What can I do? Please help.

[Sundar_7]

You have probably lost your inetd. But you should still be able to login to the console. Try again in the server's console as root.

[Harry Yu]

I did try with the ROOT on the console and it won't take it. It only show "incorrect" after I entered ROOT and password. I tried my user account and same result. My last resource is reboot the server but I doubt that'll do anything.

[Michael Tully]

If there are users still logged on, is there a chance you could use one of their sessions to investigate?

Look for inetd not running
messages in syslog like eqmemsize or anything else of note.

Rebooting the server may work, but it may make it worse. Check to make sure files haven't been removed like /etc/passwd

[Harry Yu]

No, users are locked in our database. As soon as they log off from the database, they're disconnected. I can't interrupt the disconnection.

I will try to restart the server. But the only way to do it is shut it off. I do not see any other way to do it.

[Harry Yu]

The immediate problem is I have no access to the shell at all. I can't even get pass the login, so I can't enter any commend to check the files' existance.

[D Block 2]

is there anyway, you can confirm you have a /etc/passwd file ??

do you have a backup of your /etc/passwd file?

can you try to run: rcp from another system as root ?

systemB: cd /etc; rcp passwd badsystemA:/etc

[Harry Yu]

I think it turned out one of my system drive is down. I rebooted the server and it booted me into Single User Mode and I don't see any file in usr. Do you remember how to use the ioscan to bring up the list of drive that's available?

[Michael Tully · Accepted solution]

Try:

# ioscan -fnC disk

and then use
# diskinfo /dev/rdsk/cXtYdZ

on each disk. This will basically tell you if a drive is dead or not.

[Harry Yu]

Thanks,

I can use ioscan but I can't use diskinfo since the command is not there. I think it's gone along with the system drive. When I run ioscan, I see all 10 hard drives that we have there. All of them are clam. But I can't use the diskinfo to verify them.

[Harry Yu]

I didn't mean "clam", I meant "claimed".

[Harry Yu]

When I reboot the server, I got the following errors...

lvlnboot: Warning: Couldn't query physical volume "/dev/dsk/c?t?d?": The specificed path does not correspond to physical volume attached to this volume group.
- Using current uid(0) instead.

and the disks are...
c0t0d0
c2t0d0
c0t2d0
c2t2d0

Any idea?

[Michael Tully]

what does 'vgdisplay -v' report?
look for stale lv's and the values in Cur LV and Open LV

[Harry Yu]

vgdisplay shows...
-------------------------------------------
Cannot display volume group /dev/vg0?
Volume group not activated.
-------------------------------------------

and that including vg01,vg02,vg03 and vg04.