abrams@bnlux0.bnl.gov (The Ancient Programmer) (03/11/89)
In article <978@mailrus.cc.umich.edu> wisner@mailrus.cc.umich.edu (Bill Wisner) writes: >Why not just run expire nightly with a 1 day expiry for junk? Similar >effect, no risk. Great idea. Thanks. Expire does not allow for something like: expire -n junk -e 1 -E 15 -n all -e 15 does it? Or should I fire up another expire at a time when it would not collide with the "normal" expire? Two expires, running at the same time does not sound like a great idea. -- INTERNET: abrams@bnlux0.bnl.gov BITNET: abrams@bnlux0.BITNET UUCP: ...philabs!sbcs!bnlux0!abrams
wisner@mailrus.cc.umich.edu (Bill Wisner) (03/11/89)
> Expire does not allow for something like: >expire -n junk -e 1 -E 15 -n all -e 15 > does it? Or should I fire up >another expire at a time when it would not collide with the "normal" expire? >Two expires, running at the same time does not sound like a great idea. No. At least, not in 2.11. Just create a shell script that runs two consecutive expires and execute it nightly. That way, only one cron job is created and you don't need to worry about one expire stepping upon the toes of the other.
jmdoyle@phoenix.Princeton.EDU (Jennifer Mary Doyle) (03/14/89)
In article <1075@bnlux0.bnl.gov> abrams@bnlux0.UUCP (The Ancient Programmer) wr ites: > I'm news administrator at a site with limited disk space, and >I would like to know if there are good reasons for not making the junk >directory a link to /dev/null. Will this have any other repercussions >besides loosing all of the junked articles? Where I am (Princeton University), junk is a catchall newsgroup. Local posts about conferences/speakers show up there. Recently, articles for newly created groups have shown up there, too. Ex.: Before rec.music.dementia, soc.couples, and soc.culture.turkish all became groups locally, articles posted to them from elsewhere showed up in junk. This allowed one to keep up with the groups until the newgroup took effect. Jen -- It's nice to know that when the whole world seems crazy, /\ /\ you have friends who make it seem sane in comparison. ||_____|| Jen Princeton `92 jmdoyle@phoenix.princeton.EDU | _ | Disclaimer: I am a student, I represent the future. |__( )__|
news@drivax.DRI (Wayne Davison) (03/14/89)
In article <1079@bnlux0.bnl.gov> abrams@bnlux0.UUCP (The Ancient Programmer) writes: > Expire does not allow for something like: >expire -n junk -e 1 -E 15 -n all -e 15 > does it? No, not unless you modify it, like I did. I have a ridiculously simple patch which allows you to specify a separate expire time for each group or category, as in: expire -E 15 -e 15 -n junk:1,talk:7,soc:7,comp.sys.ibm.pc:3,all This is a very simple modification to the pattern matching subroutine to set the expire time for each match. The main body of the code is left unchanged. Works like a charm, and takes as long as a "normal" expire. >Two expires, running at the same time does not sound like a great idea. You betcha, very bad news. (No pun intended, really :-) You have to run them sequentially, using a shell script. This, of course, takes twice as long to run. -- Wayne Davison ...amdahl!drivax!davison =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= Lkung tku nrcd pdnm bknwrea xztmykazojb nrcd ypdbd? Yzt "xztmyk", ypd xztmykazoj-nkwdz'b qzrdeg. Oy o xkjm.bkuzxdb.aojdb ozxprwdz edoz tku.
barnett@crdgw1.crd.ge.com (Bruce Barnett) (03/15/89)
In article <1075@bnlux0.bnl.gov> abrams@bnlux0.UUCP (The Ancient Programmer) wr >ites: > I'm news administrator at a site with limited disk space, and >I would like to know if there are good reasons for not making the junk >directory a link to /dev/null. Will this have any other repercussions >besides loosing all of the junked articles? I have always felt that if I got articles in the junk directory, then something was wrong and should be fixed. Either you are getting articles you don't want, or you are getting articles yo do want and they are not going to the right place. You can either fix your sys file, or create the newsgroups, right? -- Bruce G. Barnett barnett@ge-crd.ARPA, barnett@steinmetz.ge.com uunet!steinmetz!barnett
cks@ziebmef.uucp (Chris Siebenmann) (03/19/89)
I can only see two problems with this:
1. you'll never notice newsgroups whose creation your site has missed
2. if an article comes in that winds up in junk and some other groups,
the news system could decide to queue it for downstream feeds as
junk/<whatever> (which wouldn't exist) as opposed to the article
number in a newsgroup you keep around.
[this may be a Cnews-ism, and is perhaps a problem in the news software.]
However, doing this probably won't gain you much, and what you do
gain can be gained in a simpler manner. Consider why articles wind up
in junk:
1. you don't have any of the newsgroups they're in
a. because you missed the newgroup message for the groups, in which
case you want the article anyways and should create the newsgroup
b. they're in newsgroups you don't want at all, in which case you
should tell your upstream feed to turn them off and save
modem time.
2. they're too old, in which case you may want to figure out *why* and may
want to see them anyways.
3. they're crossposted between some groups you do get and some you
don't. Here deleting the article in junk buys you no gain in disk
space, because it will just stay around in the other groups.
On a well-configured news site, anything except #3 should be very
rare and easily handled by hand.
--
"Though you may disappear, you're not forgotten here
And I will say to you, I will do what I can do"
Chris Siebenmann uunet!{utgpu!moore,attcan!telly}!ziebmef!cks
cks@ziebmef.UUCP or .....!utgpu!{,ontmoh!,ncrcan!brambo!}cksromain@pyramid.pyramid.com (Romain Kang) (03/20/89)
Junk can be caused by things that cannot be fixed. Can you say "inet leakage"? pyrnj gets about 600-700KB of inet leakage from rutgers during an average week. Remember, rutgers is a well-administered site by Usenet standards, run by diligent and knowledgeable staff. (It makes me worry to think what promoting all the inet newsgroups to full Usenet might do to countless /usr/spool's around Usenet.) Here's something from pyramid's sys file, to serve as guardian of the corporate decorum. You can't do this for all the inet newsgroups -- the list is too long to fit LBUFLEN. junk:junk,alt.birthright,alt.california,alt.flame,alt.individualism,\ alt.rhode_island,alt.rock-n-roll,alt.sex,talk.bizarre:U:rm -f /tmp/%s %s pyrnj handles inet leakage by aliasing all of those groups to the "inet." prepended to the group name. Before every sendbatch, a script runs to remove the prefix from newly arrived inet articles, preventing bogus group names from going back into Usenet if an inet newsgroup has been cross-posted with regular Usenet. This way I can isolate the inet newsgroups, and conceivably feed them downstream of pyrnj if desired, without excess inet propagation. Crufty? To the max! But no one's complained since I starting repairing the newsgroup names. -- Now for a version of expire that runs on the kitchen memo board...
abrams@max.arpa (The Ancient Programmer) (03/22/89)
In article <45@crdgw1.crd.ge.com> barnett@crdgw1.crd.ge.com (Bruce Barnett) writes: >In article <1075@bnlux0.bnl.gov> abrams@bnlux0.UUCP (The Ancient Programmer) wr >>ites: >>I would like to know if there are good reasons for not making the junk >>directory a link to /dev/null. Will this have any other repercussions >>besides loosing all of the junked articles? > >I have always felt that if I got articles in the junk directory, then >something was wrong and should be fixed. > >Either you are getting articles you don't want, or you are getting (^^^^^^^^^^^^^^you bet) >articles you do want and they are not going to the right place. > >You can either fix your sys file, or create the newsgroups, right? ^^^^^^^^^^^^^^^^^ Point well taken, however these junked articles are in newsgroups which are in my sys file with the ! preceding them. I assume that's why they are being junked. Many are for distributions such as wny(west new york?), or buf(buffalo?) and others. They are apparently leaking all over my junk directory.;-) INTERNET: abrams@bnlux0.bnl.gov BITNET: abrams@bnlux0.BITNET UUCP: ...philabs!sbcs!bnlux0!abrams
barnett@crdgw1.crd.ge.com (Bruce Barnett) (03/23/89)
In article <1087@bnlux0.bnl.gov>, abrams@max (The Ancient Programmer) writes: > Point well taken, however these junked articles are in newsgroups which >are in my sys file with the ! preceding them. I assume that's why they are >being junked. Many are for distributions such as wny(west new york?), or >buf(buffalo?) and others. They are apparently leaking all over my junk >directory.;-) I suggest you tell your feeds to put a !wny and !buf if their sys file for your machine. Why should you pay for unwanted junk mail that someone insists on sending you? -- Bruce G. Barnett <barnett@crdgw1.ge.com> a.k.a. <barnett@[192.35.44.4]> uunet!steinmetz!barnett, <barnett@steinmetz.ge.com>