molenda@s1.msi.umn.edu (Jason Molenda) (05/06/91)
I was just checking my history file for duplicate message ids with the following line a little earlier: awk '{print $1}' < /usr/lib/news/history | tr '[A-Z]' '[a-z]' | sort | uniq -d >/tmp/boog to see how significant case was in Message-IDs (I was about to hack up a piece of news sotware to make it case sensitive temporarily). I turned up something rather odd, though (the last few lines of /tmp/boog, there were about 900 lines in the file on my 15-day history file): <uc7s6ck00awiqokyqi@andrew.cmu.edu> <uc7uchc00uzx44f84x@andrew.cmu.edu> <wc7sjfm00wb9aiem4i@andrew.cmu.edu> <wc7sngm00wb9eiendi@andrew.cmu.edu> <wc7ttro00uh744rnqz@andrew.cmu.edu> <yc7rw0o00waxmwieml@andrew.cmu.edu> <yzz911w164w@cellar.uucp> So if I check the last one with the following command: egrep -i 'yzz911w164w@cellar.uucp' /usr/lib/news/history I get <yZZ911w164w@cellar.UUCP> 673077310~- (Wed May 1 00:55:10 1991 CST) <yZZ911w164w@cellar.UUCP> 673090384~- (Wed May 1 04:33:04 1991 CST) with the human-readable stuff in parens added by me. Why is this? The message ids are, as far as I can tell, identical. Is this some weird thing that happens with cancelled articles (I don't know if it was a cancelled article or if it just expired; I only have about 150 megs for news so chances are it was expired already). I'm running the March 24 Cnews on a Sun 3/60 running SunOS 4.0.3 with DBZ and NNTP for my news transport mechanism. Am I missing something obvious here? -- Jason Molenda, Tech Support, Iris & News Admin, Minnesota Supercomputer Inst molenda@msi.umn.edu || "You can tune a piano but you can't tuna fish."
henry@zoo.toronto.edu (Henry Spencer) (05/07/91)
In article <1991May6.085649.19174@s1.msi.umn.edu> molenda@s1.msi.umn.edu (Jason Molenda) writes: ><yZZ911w164w@cellar.UUCP> 673077310~- (Wed May 1 00:55:10 1991 CST) ><yZZ911w164w@cellar.UUCP> 673090384~- (Wed May 1 04:33:04 1991 CST) > >with the human-readable stuff in parens added by me. > >Why is this? The message ids are, as far as I can tell, identical. Is >this some weird thing that happens with cancelled articles... Nope. Two entries for the same message ID means something's wrong with duplicate rejection. No specific thoughts on what, although the standard caution is to make certain that relaynews and expire agree on what package is being used for the index, i.e. that they got rebuilt at the same time with the same library. -- And the bean-counter replied, | Henry Spencer @ U of Toronto Zoology "beans are more important". | henry@zoo.toronto.edu utzoo!henry