when configured to use ruptime, but when the client machines arn't running the ruptime daemon, nanny failed the connections but apparently doesn't close the file handle. Eventually (after afew hours) nanny runs out and dies.
Thank you for this report! We will try to fixed this in an upcoming errata
Hi, I also experienced the problem. The following seemed to have fixed it for me. The nanny version is the one that comes with piranha-0.4.17-7. --- nanny.c.orig Wed Jan 24 10:53:17 2001 +++ nanny.c Thu Jan 25 08:58:53 2001 @@ -750,12 +750,14 @@ logArgv(flags, argv); kill(child, SIGKILL); waitpid(child, &status, 0); + close(p[0]); return NULL; } if (!WIFEXITED(status) || WEXITSTATUS(status)) { piranha_log(flags, (char *) "The following exited abnormally:"); logArgv(flags, argv); + close(p[0]); return NULL; } @@ -781,7 +783,7 @@ *argp++ = loadCommand; *argp++ = inet_ntoa(*remoteAddr); - if (strcmp(loadCommand, (char *) "none")) + if (strcmp(loadCommand, (char *) "none") == 0) return -1.0; /* nothing to see here, go home silently */ if (strcmp(loadCommand, (char *) "rup")) Cheers Tinus
Yeap, that looks sane. I'll add that into the cvs shortly. If I get a moment to myself I'l make the rpms as well (bit busy stomping on the kernel the last while) Phil =--=
These patches (two "close(p[0])" statements and a " == 0") have been added to nanny and will appear in an upcoming release, with credit given to tinus. Users can apply the above patch as a work-around until the new nanny has been released.
*** Bug 53783 has been marked as a duplicate of this bug. ***