> There was a specific problem with exit which was fixed by Ted Tso's patch.
> There _should_ be no other such problems, except perhaps users intentionally
> leaving things running (or processes that do it themselves).
Yes. Put that one into 1.2.13 (after some hacking). Now we are running
1.3.57 (to get workable quotas). No change in the problem. With the
exception that ntalkd's are now hanging in massive quantities. I
circumvented that problem by turning the service off and will tackle a
permanent fix later. :-)
>
> The other report I saw here (about needing ^Q) is indicating what is
> probably a telnetd problem not resetting the terminal state cleanly.
>
> Are these processes attached over tty/pty pairs, do their stdin/stdout point
> to stuck tty/pty pairs, what telnetd are you using ?
>
Yes, for us, all processes are pty. Yes, they they do seem to point that
way. We have no user-accessible direct serial connections to the cluster.
We use the standard telnetd that comes with Slackware. The source is from
NetKit-B-0.05. Compiled as both ELF and a.out (depending on the node. The
result is the same problem on either).
-Jonathan _ _
------------------------------------------------------------->>>>>>>>-(o)(o)---
Jonathan A. Davis | Academic Systems Analyst | Hattiesburg/Gulf Park/Stennis
Computing Center | Box 5171 | 39401-5171 | (601) 266-4103 | davis@cc.usm.edu
http://evergreen.cc.usm.edu/~davis | Linux: The choice of a GNU generation