Summary: proc: table is full

From: K. Bader (
Date: Fri Jan 04 1991 - 15:58:37 CST

Sun managers,

I got a lot of responses to my question on why I was getting the message
*proc: tale is full* echoed to the console every few seconds before I sent
my "nevermind" letter in. Nonetheless, I received valuable information
which I will share especially since it was requested.

My problem was I had lpd's runnning out of control because my printcap
file was corrupted on an interleaf installation. Generally though, if this
occurs you can do a /etc/pstat -T to see what your process table looks like.
The fix is either increasing MAXUSERS and reconfiguring the kernal or editing
the file param.c to increase the value of NPROC.


more details

From: (Tim Ramsey)

Try seeing what "pstat -T" tells you. This will give you the number of
process table entries used/total available.

You can increase this without increasing maxusers by editing

(where {karch} is sun3, sun3x, sun4, or sun4c and {kernel_name} is the
name of the kernel generated by /etc/config, should be given in the first
line of /etc/motd).

From: Doug Peterson <>
Check the output from pstat -T.

pstat -T
167/364 files
158/158 inodes
 35/ 74 processes
 10/ 32 files
4840/16264 swap

While processes (nproc) is a function of maxusers, it's actually computed in

#ifdef sun
#define NPROC (10 + 16 * MAXUSERS)
#undef MAXUPRC
#define MAXUPRC (NPROC - 5)
#define NPROC (10 + 8 * MAXUSERS)
int nproc = NPROC;
int maxuprc = MAXUPRC;
int ninode = (NPROC + 16 + MAXUSERS) + 64;
#ifdef sun3x
int minpts = 16 * MAXUSERS;

You might want to cut back on MAXUSERS and tweak the contstant.

Thanks to:
From: liz@cgdisis.cgd.ucar.EDU (Liz Coolbaugh)
From: era@niwot.scd.ucar.EDU (Ed Arnold)
From: (Henry Clark)
From: fed!m1rcd00@uunet.UU.NET
From: Steve Hanson <>
From: (Tim Ramsey)
From: auspex!guy@uunet.UU.NET (Guy Harris)
From: Doug Peterson <>
From: Barbara Fraser <byf@SEI.CMU.EDU>

This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:06:10 CDT