[ale] SSH session ends immediately after authentication

[Dow_Hurst]

Jason,
One thought is that all available processes were full and no other ttys could start.  If you have a IDE drive you could run smartd and check it that way for occurring errors.  Two thoughts there...
Dow


-----Original Message-----
>From: Jason Day <jasonday at worldnet.att.net>
>Sent: Jan 27, 2006 9:40 AM
>To: Atlanta Linux Enthusiasts <ale at ale.org>
>Subject: Re: [ale] SSH session ends immediately after authentication
>
>On Thu, Jan 26, 2006 at 03:42:47PM -0500, Jason Day wrote:
>> On 1/26/06, James P. Kinney III <jkinney at localnetsolutions.com> wrote:
>> 
>> > Ah Ha! A full hard drive would cause a login to fail as it can't record
>> > the access. If this is a production server you have a problem as the
>> > logging is now not running.
>> >
>> 
>> No, it's just my cable modem router for home use.  The downtime isn't a big
>> deal, but a compromised system would be.  This makes me slightly less
>> anxious.  I'll report what I find after I get home.
>
>Well, it wasn't a full hard drive.  When I got home I tried logging in
>on a serial console without success.  So I dragged a monitor over and
>plugged it in and still couldn't login.  I did however have an open ssh
>session connected from another box on my LAN.  df showed plenty of space
>on the hard drive, netstat didn't show anything unusual.  Tried to su so
>I could check the logs, and su segfaulted.  I tried another command
>(don't remember which one) and the ssh session closed without warning.
>
>Rebooted with a knoppix disk and ran memtest86.  No errors.  Tried
>rebooting again, but I kept getting IO errors when knoppix was trying to
>mount the cdrom.  I think the cdrom drive might be bad, it is ancient.
>Rebooted to single user mode, everything seemed fine.  Nothing unusual
>in the logs.  The logs show all my ssh attempts, but don't show any
>errors; just a session opened immediately followed by a session closed.
>During all this time, fetchmail continued to run *and* successfully
>fetched mail, and exim and procmail successfully delivered it.
>
>I don't know what would make it go all wonky like that, but it occurs to
>me that the programs that had problems (su, login, ??) are all on the
>root partition, while the programs that worked (fetchmail, sshd, exim,
>procmail) are on the /usr partition.  I'll get a toms root boot tonight
>so I can do a badblocks check on the root partition.  Anything else I
>should look out for?
>
>Thanks,
>Jason
>-- 
>Jason Day                                       jasonday at
>http://jasonday.home.att.net                    worldnet dot att dot net
> 
>"Of course I'm paranoid, everyone is trying to kill me."
>    -- Weyoun-6, Star Trek: Deep Space 9
>_______________________________________________
>Ale mailing list
>Ale at ale.org
>http://www.ale.org/mailman/listinfo/ale


No sig.