in.telnetd and in.ftpd problems

[tckirk]

I am runing Solaris 9 on a E-450. users can not telnet or ftp into the system after it is online for several days. Rebooting the system fixes the problem for several days.
What can cause telnet and ftp services to stop responding ?

 

[vptl]

Fot this type of issues you need to trbshoot.
when user cannot login ...go on console and see if you can ps -ef | grep inet
see if you can ls -l /usr/lib/in.telnet or in.ftp.
Basically...three thing you need.
1. in.telnet accessible.
2. inetd is running.
3. inetd.conf have info about in.telnet.(In your case it is there b/c it works after reboot.)

 

[ananthkrish]

Due to killing sum of the background processes in server , wii cause this problem . without killing or stoping the telnet & ftp services it wont stop.

 

[tckirk]

Loging in at the Console is not an option when this happens. When I attempt to login in at the console the console flashes back to the CDE Welcome screen again after I enter my password and select OK at the CDE welcome screen.
I am still looking for clues as to where to look for the problem.

 

[coffeysm]

I would check your .dt, .Xauthority, etc... under your home directory if CDE kicks you back out. Also check the permissions on the location where /dev/null is linked, I have seen instances where /devices/pseudo/mm@0:null has permissions set to 440, it should be 666.

Are you running any kind of wrapper programs for Telnet and FTP. Also next time it stops responding, instead of rebooting I would try snooping the ports, running truss -p on the process id number, and checking /var/adm/messages for any clues...

 

[spamly]

I'm guessing it is a inetd problem. Make sure you have the appropriate telnet and ftp lines in your /etc/inetd.conf. Also check to see if there is a inetd patch for Solaris 9 available.

 

[SamBones]

You aren't filling up [tt]/tmp[/tt] are you? I've seen that cause odd behaviour like this. Especially if a reboot fixes it.

Make sure nothing's messing with [tt]/tmp[/tt], [tt]/proc[/tt], and that you have plenty of free swap space. Also if you don't have [tt]/var[/tt] on its own file system, make sure the root partition isn't filling or full.

Like coffeysm says, check [tt]/var/adm/messages[/tt] for anything suspicious.

Is the time it locks consistent? If the amount of time from reboot till it locks consistent?

 

[marrow]

I agree with Sam, you maybe running out of resources. Might be an idea to run "vmstat 2 5" after a reboot and then again at intervals leading upto your problem, vmstat will highlight memory & tmp space problems. Perhaps we can view results.