Path: utzoo!news-server.csri.toronto.edu!rutgers!usc!rpi!uupsi!unvax.union.edu!boehme
From: boehme@unvax.union.edu (Eric M. Boehm)
Newsgroups: comp.unix.ultrix
Subject: Re: gnode: table is full
Message-ID: <1991Mar12.015319.22984@unvax.union.edu>
Date: 12 Mar 91 01:53:19 GMT
References: <6081@amc-gw.amc.com> <2663@shodha.enet.dec.com> <19682@cbmvax.commodore.com> <2671@shodha.enet.dec.com>
Distribution: usa
Organization: Union College, Schenectady, NY
Lines: 27

alan@shodha.enet.dec.com ( Alan's Home for Wayward Notes File.) writes:

>In article <19682@cbmvax.commodore.com>, grr@cbmvax.commodore.com (George Robbins) writes:
>> [ George reasonably points out that the message "out of gnodes" can
>>   can come from a disk that's out of gnodes, as opposed to the kernel
>>   table running out. ]

>	I haven't seen it on a V4.x system, but sometimes on a V3.x
>	systems a disk that goes offline would eventually cause the
>	"out of gnodes" message to be printed.  So, broken hardware
>	can also cause it.

I have seen this -- specifically on the filesystem where news is
stored. I have rebuilt the filesystem with new parameters to increase
the number of inodes but I have still seen it when there is a lot of
news. This was under 3.x, 4.x. I increased the number of maxusers but
I think it could still happen.

As illustration, here is the report from df -i:

Filesystem   Total    kbytes   kbytes   %     inodes   inodes   %            
node         kbytes   used     free     used  used     free     used Mounted on
/dev/rz0e     125015   77712   34801    69%    25047    7721    76%  /var/spool
-- 
Eric M. Boehm
boehme@unvax.Union.EDU
BOEHME@UNION.BITNET