Telnet/FTP hangs on SGI O2 IRIX 6.5

>but it never prompts you for a username/password

Have you tried waiting until the remote side actually closes the connection?  Sometimes a wait of "forever" just means "I got tired of waiting" rather than "I received a timeout error", which is the only reason I ask...

Deploy a sniffer - it should help narrow your focus.

Cheers,
-Jon

If it's just slow at giving a login prompt (up to 2 minutes), it may be a DNS issue - The server can't do a reverse lookup on the incoming client connection.
https://www.experts-exchange.com/questions/20668120/FTP-on-SCO-Unix-slow-to-connect.html

If it _never_ responds, it may be that the server can't spawn a login process, e.g. if there are too many processes already running. That should generate a error message in /var/adm/SYSLOG

Right tfewster.... another thing I've on occasion (waaay back:-) have seen is that the allocation of a pseudo-TTY fails. I've observed that mostly on older Suns (SunOS was cr*p in this regard). Usually produced something slightly unintelligible in the syslog, of course, but on some systems just looked like a forever wait. A restart would (of course) solve that, but one could also futz with the ptys a bit ("unwedge" them manually:-).

It might (as has been mentioned) be that the system is really busy too... When it is unresponsive, does it barf anything on the system console? Is it possible to log in locally?

Cheers
-- Glenn

thanks for all your suggestions. The problem has not cropped up again so I can't try anything you suggested right now. I've past along the information to my users. Thanks for your assistance.

Alright, here is what is displayed -

"ALERT: pmcd[742] - out of logical swap space during brk/sbrk - see swap[1M]", "critical memory error"

Computer then requires a reboot to restore telnet/ftp functionality.

It looks like you have an underlying problem with the machine running out of virtual memory (RAM + swap), so there is no free memory to spawn a login process for telnet/ftp.

I presume you increased the swap space after discussing this with ahoffman in
https://www.experts-exchange.com/questions/23958507/ALERT-pmcd-804-out-of-logical-swap-space-during-brk-sbrk-see-swap-1M-Error-message.html
as the problem isn't happening every day now?

There's an interesting article regarding pmcd at http://seclists.org/bugtraq/2000/Apr/0045.html
So the performance monitoring tool,  pmcd,  may be the culprit that is using up all your memory. Or it might just be another symptom of memory shortage

On the bright side, if pmcd is running, you should be able to use it to analyse the system performance and show you where the problem lies :-)   Otherwise, running basic tools like ps/vmstat/sar regularly would enable you to see if memory usage is growing over time and pin down the real culprit

Since I'm on vacation (and hence late to the party...:-) all I can add is my CC to that Tim.

Cheers
-- Glenn

Back from vacation, and on to a 4-day (full weekend) "conference" in Abisko ... "Survival" and "The great outdoors" coupled with some beverage intake = Sauna+really cold bath in Torne Träsk (no, not really a swamp, just a big lake far above the polar circle) and a really heavy head... So no, no brilliant insights;-). But you've nailed the symptoms and likely cure (simply increase the virtual memory). One could probably look long and hard at memory usage/process, to see if this is a memory leak or something else (just a "high load" kind of thing (no, not load as in load average;-)). But you've covered that nicely too.
As to why on EE when on vacation... It was a long one, I was bored by all the manual labour (renovation etc):-):-).

Cheers (eh, well... I think I'll have a "white week" or two, so not really:-)
-- Glenn