thisted@gargoyle.UChicago.EDU (Ronald A. Thisted) (12/11/87)
We have a small network of SUN-2s: a file server and two diskless nodes. Occasionally one of the diskless nodes starts going into a frenzy of spawning inetd processes, at the rate of about 5 per second. The load average shoots up and stays there until someone notices and kills the parent daemon and restarts inetd. We have not been able to replicate the problem; it seems to recur apparently at random about once or twice a week. One annoyance is that the accounting file rapidly fills up with records of the inetd's. It _appears_ as if the biod daemons on the server are also abnormally active whenever the inetd problem is occurring on the clients. A similar problem is that from time to time our server's nfsd daemons seem to go crazy, absorbing among them about 50% of all cycles, regardless of apparent activity elsewhere on our local net. I'm not a unix hacker, but would appreciate any clues or suggestions as to what we can do. I don't generally read this group, so mail is preferred. If others have had the same problem, send me mail and I will send you a summary of what I learn. If there is enough interest I will post, too. Ron Thisted Dept of Statistics The University of Chicago Internet: thisted@galton.uchicago.edu UUCP: ...!ihnp4!gargoyle!galton!thisted Bitnet: thisted%galton@UChicago