I have users at several remote locations that connect to a AIX 5.1 server using telnet. Last week the users at one location reported problems with logging into the server. After entering user ID and password the users get a blank screen. I have checked the user password, profile, and permissions all is in order. It is not just one user it is all users at this location. I have the users connect to Citrix or VPN they are able to login in using their user ID and password. This location has been live for over six months. All other locations have no issues some have been on line for several years. The server has been in place for about four years. Where else can I look problems?
Does it take about 3 minutes, and then finally login?
Does this only happen to some client machines?
Get the IP address of the client who is trying to login then confirm you can do a reverse name lookup of their machine using "nslookup".
Yup, once a user logs in (successful authentication) then this is logged to wtmp:
# /usr/sbin/acct/fwtmp </var/adm/wtmp | more
gus pts/0 pts/0 7 471118 0000 0000 1169145986 machine.hostname.com Thu Jan 18 12:46:26 CST 2007
If there is a problem with reverse-resolving the IP to a hostname, then the login process will appear to hang until the problem is resolved (or until there is a timeout). I would guess that the authoritative DNS server for the remote site is down or not reachable from the primary site. You can prove this theory by adding the remote system(s) to /etc/hosts, but you must ensure that name-resolution order uses the hosts file first:
/etc/netsvc.conf:
hosts=local,bind
Thanks for responding
porter
The users never get logged in the session times out.
The problem is with all machines at this location.
Reverse name lookup did not work because the domain for this location was not added. I have had the domain added to DNS and now reverse name lookup is working.
gus2000
The user information dose appear in wtmp when user attempts to login.
I did add the workstation to the host table but was still unable to login.
The users are still unable to login.
The following summarizes the AIX login process:
* A user enters an id at the login prompt.
* getty starts the login program to check /etc/passwd and /etc/security/passwd.
* login starts a login shell.
Since the attempt is logged, I'd say it's likely an authentication problem. What is the authenication method for these users? Is it different from the other users? And what is the shell for these users?
If it is an authentication problem, then the failed logins should be listed here:
/usr/sbin/acct/fwtmp < /etc/security/failedlogin
It is possible that the logiles utmp, wtmp and failedlogin could impede the login process if corrupted, but that would tend to be a systemic problem and not just for a subset of users. Still, since the latter two logs grow without limit, this might be a good time to archive and clear them.
You might also want to check the output of "lslicense". The number of simultaneous logged-in users is limited to this number (although you can change it with "smit chlicense").
The authentication method is the same for all users.
All users use Korn shell.
There is not an entry in /etc/security/failedlogin for the user I tested with this morning. I do have a script setup to clean the wtmp file and now that I know about the failedlogin file I will set up a script to clean it.
The License limit is set to 500. I have about 150 to 170 users connected to the server at any given time.
This problem was caused by a network equipment failure. The equipment has been replaced and all users are able to login.