randy@ux1.cso.uiuc.edu (01/16/90)
We've had a problem with lookupd for a while now. I don't know when it started, but at one point lookupd just acquired an insatiable appetite for memory. The virtual memory allocated for it (as indicated by ps) just grows (roughly linearly) until our file system fills up (660 Meg HD)! I would say the rate of growth is on the order of 20 Meg/day. Needless to say, we periodically have to reboot it when we begin running short of space (about once every two days or three days now). Of course, this is the swapfile that's growing - as lookupd demands more and more virtual memory. I don't have a single solitary lead as to how to fix this (or even where to go for info - as lookupd is NeXT specific). AAAAAAH! Please, can someone help us? Our NeXT is crippled!
alg@cornell.UUCP (Anne Louise Gockel) (01/17/90)
In article <18600003@ux1.cso.uiuc.edu> randy@ux1.cso.uiuc.edu writes: > >We've had a problem with lookupd for a while now. I don't know when it >started, but at one point lookupd just acquired an insatiable appetite >for memory. The virtual memory allocated for it (as indicated by ps) >just grows (roughly linearly) until our file system fills up (660 Meg >HD)! I would say the rate of growth is on the order of 20 Meg/day. >Needless to say, we periodically have to reboot it when we begin >running short of space (about once every two days or three days now). >..... Shortly after 1.0 came out, NeXT told us that there is a bug in the interaction between yp and lookupd. Apparently the only solution is: - niload /etc/group into NetInfo. Do not niload the yp "magic cookie" - completely remove the /etc/group file - do not use yp to serve groups Of course, this is extremely ackward in a distributed, predominantly YP, environment. NeXT did not say when they planned to fix this serious bug. I certainly hope it will be fixed in the next release as it is possible, but rather ackward, to work around. If you are not running YP, then I do not have any idea what is causing your memory leak. It sure sounds like the one we had that was YP related. -Anne Louise Gockel Cornell Computer Science Internet: alg@cs.cornell.edu UUCP: cornell!alg Bitnet:ANNE@CRNLVAX5