Addendum to: SUMMARY: CHILD_MAX & osf/1 v3.0

From: <haymanR_at_icefog.uacn.alaska.edu>
Date: Fri, 23 Jun 1995 08:23:13 -0800

From: "Dr. Tom Blinn, 603-881-0646" <tpb_at_zk3.dec.com>

|>Randy, I looked at <sys/syslimits.h> and param.c and getconf.c and sysconf.c
|>and it looks like what's happening is that the value returned by getconf is
|>a constant:
|>
<code snipped>
|>
|>What happens is that getconf calls sysconf with a function code of
|>_SC_CHILD_MAX and sysconf() returns the value of CHILD_MAX that was defined
|>in <sys/syslimits.h> AT THE TIME THAT libc WAS BUILT.
|>
|>What you have to look at is maxuprc in the running kernel. You probably can
|>do this with kdbx (but I don't remember exactly how to do it).
|>
|>Note also that the maximum THREADS per user is dependent on CHILD_MAX, not
|>on MAXUPRC (see param.c) so if the problem you're seeing is running out of
|>threads, not processes, you need to change maxuthreads (by changing
|>CHILD_MAX) as well as, perhaps, MAXUPRC.

This clears up the v3.0 and v3.2 differences (in my mind).

Thanks to all who responded.


Randy M. Hayman
haymanr_at_icefog.alaska.edu
Received on Fri Jun 23 1995 - 19:11:19 NZST

This archive was generated by hypermail 2.4.0 : Wed Nov 08 2023 - 11:53:45 NZDT