brianop@salem1.UUCP (Brian McBee) (12/03/88)
I'm doing an informal poll to see how much news is kept on all those sites out there. Write me and tell me how many days worth of news you generally keep on your system, and/or how many megabytes you dedicate to news storage. If anyone's interested, I will post the results. Thank you. -- -------------------------------------------------------------------------- What the eye beholds CI$: 72406.1363 And the heart covets PLINK: Brianop Let the hand boldly sieze! UUCP: ...tektronix!tessi!agora!salem1!brianop --------------------------------------------------------------------------
karl@triceratops.cis.ohio-state.edu (Karl Kleinpaste) (12/07/88)
brianop@salem1.UUCP (Brian McBee) writes:
I'm doing an informal poll to see how much news is kept on all those sites
out there.
I asked essentially the same question about a year ago, and got lots
of interesting responses. Space dedication varies mighty wildly from
place to place. Here, we use an entire Eagle drive, and expire weekly
on a 6-week basis, on Saturday early AM:
[200] [10:43am] tut:/dino0/karl# df /usr/spool/news
Filesystem kbytes used avail capacity Mounted on
/dev/iop/pdisk03h 386942 326442 60500 84% /usr/spool/news
--Karlaad@stpstn.UUCP (Anthony A. Datri) (12/08/88)
In article <KARL.88Dec6110554@triceratops.cis.ohio-state.edu> karl@triceratops.cis.ohio-state.edu (Karl Kleinpaste) writes: >Here, we use an entire Eagle drive, and expire weekly >on a 6-week basis, on Saturday early AM: >[200] [10:43am] tut:/dino0/karl# df /usr/spool/news >/dev/iop/pdisk03h 386942 326442 60500 84% /usr/spool/news That's fairly incredible!:-) When I arrived here, news was kept in a recycled nd partition (Sun 3/180). If you don't know what nd partitions are, be very glad. This one was around 20 meg, and loved to run out of space. When it ran out of space, news would go on creating 0 length files, taking up all the inodes. Then when I got rid of the junked messages that were taking all the space, I still didn't have any inodes and had to find down the tree and nuke 0 length files. I solved our problems incrementally: I changed news feeds, from a very clumpy (no news for 5 days, then WHAM!) 1200 baud feed to a very uniform TrailBlazer feed. I changed the feed's sys entry for us to list only the groups we wanted, with a couple of wildcards like news.*. Voila! No more junking. I also repartitioned our disks in a sane manner -- 10 partitions on a disk with 2 meg free in each are nowhere *near* as useful as 1 partition with 20 meg. So I made as few partitions as I could get away with with SunOS 3.2's idea of being a server and without concentrating activity on one disk, and ended up with a big partition where I put all of /usr/spool (including news), /usr/local stuff, and other stuff that used to be recycled nd partitions (yuck). So now it stands with plenty of room and plenty of inodes to go around: <=>5<=>df /muppet Filesystem kbytes used avail capacity Mounted on /dev/xy0d 217871 150599 45484 77% /muppet <=>6<=>df -i /muppet Filesystem iused ifree %iused Mounted on /dev/xy0d 17134 46354 27% /muppet News used to be expired every night on a 7/21 day basis; I only yesterday changed that to 10/21 to reflect our newfound happiness. We get a moderate number of groups, and our /usr/spool/news adds up to about 18 meg right now. -- @disclaimer(Any concepts or opinions above are entirely mine, not those of my employer, my GIGI, my VT05, or my 11/34) beak is@>beak is not Anthony A. Datri @SysAdmin(Stepstone Corporation) aad@stepstone.com stpstn!aad