Path: utzoo!utgpu!water!watmath!clyde!rutgers!ames!oliveb!jerry
From: jerry@oliveb.olivetti.com (Jerry Aguirre)
Newsgroups: news.admin
Subject: Re: How long are you keeping news these days?
Message-ID: <12312@oliveb.olivetti.com>
Date: 6 Jan 88 21:38:15 GMT
References: <3899@tut.cis.ohio-state.edu> <22409@ucbvax.BERKELEY.EDU>
Reply-To: jerry@oliveb.UUCP (Jerry Aguirre)
Organization: Olivetti ATC; Cupertino, Ca
Lines: 50

In article <22409@ucbvax.BERKELEY.EDU> fair@ucbarpa.Berkeley.EDU (Erik E. Fair) writes:
>Disk space has been a problem on ucbvax last fall for the first
>time in years. We run netnews (both the spool and lib directory)
>out of an Eagle's "h" partition, which is 140Mbytes.

Interesting, my /usr/spool/news is also 1h of an Eagle.

>Then we ran out of inodes again, so I cut the expire time back to
>25 days, and the daily script mails me a "df -i" every night after
>expire completes, so that I can keep an eye on it.

With the variations in volume of news I was also having problems coming
up with with an expire time that made full use of the disk without
running over.  The problem with such a "hand" solution is that it uses
my time and if I am unavailable then things break down.  

I finally came up with a script that maintains things without attention
from me.  Using df and awk the script checks the available blocks and
inodes on /usr/spool/news.  (Somewhat like what patch 14 adds to
sendbatch.)  When this drops below a specified level (10K blocks and 3K
inodes) news is expired at 28 days.  If there is still not enough space
news is expired again at 21 days and mail is sent to me about it.  This
is repeated with shorter expire times until enough blocks and inodes are
free.

In practice expire only runs every 4 days or so.  Due to the unevenness
of the flow sometimes this is not enough and the 21 day expire will run.

>Here's this morning's report:

>Filesystem    kbytes    used   avail capacity iused   ifree  %iused  Mounted on
>/dev/hp1h     140564   89593   36914    71%   27576   11336    71%   /usr/spool/news
Mine looks like:
Filesystem    kbytes    used   avail capacity iused   ifree  %iused  Mounted on
 /dev/hp1h     115311   92688   11091    89%   35405   21939    62%   /usr/spool/news

>P.S.	Yes, I know, I could backup the filesystem and mkfs with more
>	inodes, but that would eventually leave me with disk space
>	problem instead of lack of inodes - I can basically take
>	my pick on this one, AND ucbvax would have to come down to
>	single user mode for a while we do that...

I ran into the problem of too few inodes and found out that mkfs was
creating fewer inodes than the default.  It seems there is a hard limit
on inodes/cylinder-group and specifying more won't help.  On an eagle
this hard limit is less than the default!  I finally rebuilt
/usr/spool/news with a cylinder-group size of 8 and that fixed the
problem.  Of course waiting hours for restore to finish was a pain.

					Jerry Aguirre