Yes; use supervise from the daemontools package. It's available from <ftp://koobera.math.uic.edu/pub/software/>. That will at least restart nntpcache after it dies. There are an awful lot of segv's to fix :-) Evan On Tue, 28 Jul 1998, Mark Mayo wrote: > > Hi all. I just put nntpcache-2.3.2.1 on a Solaris 2.6 box (sparc). > I'm seeing a bunch of SEGV's that concern me.. 3 of them are just > in children processes, and recovery is perfect, but once the parent > process SEGV'ed leaving me refusing connections of course.. > > If anyone has any ideas on how to fix these, please let me know > ASAP. nntpcache was compiled with gcc 2.7.2.1 > > (I apologize in advance for the long, unwrapped lines to follow) > > > Case 1: > ------- > > Jul 28 08:48:13 michelob nntpcached[352]: nntpcache.c:400: starting client task > Jul 28 08:48:13 michelob nntpcache-client[16677]: client task awakening > Jul 28 08:48:14 michelob nntpcache-client[16677]: spc-isp-van-uas-01-29.sprint.ca connect from unknown@spc-isp-van-uas-01-29.sprint.ca (209.103.0.30) > Jul 28 08:48:14 michelob nntpcache-client[16677]: sockets.c:427: <- GROUP alt.binaries.pictures.erotica.amateur > Jul 28 08:48:15 michelob nntpcache-client[16677]: sockets.c:427: <- LISTGROUP alt.binaries.pictures.erotica.amateur > Jul 28 08:48:15 michelob nntpcache-client[16677]: sockets.c:153: server attach bypassed (server tagged down) 297 seconds remaining before retry > Jul 28 08:48:15 michelob nntpcache-client[16677]: nntpcache.c:208:Connection refused: SIGSEGV! > Jul 28 08:48:15 michelob nntpcache-client[16677]: spc-isp-van-uas-01-29.sprint.ca exit articles 0 group 1 > Jul 28 08:48:15 michelob nntpcache-client[16677]: spc-isp-van-uas-01-29.sprint.ca times user 0.00 system 0.00 elapsed 2.00 > Jul 28 08:48:15 michelob nntpcached[16677]: clean shutdown with error 1. dumping core for debug analysis > > > Jul 28 08:50:53 michelob nntpcache-client[16709]: sockets.c:427: <- LISTGROUP alt.binaries.pictures.erotica.amateur.female > Jul 28 08:50:53 michelob nntpcache-client[16709]: sockets.c:153: server attach bypassed (server tagged down) 139 seconds remaining before retry > Jul 28 08:50:53 michelob nntpcache-client[16709]: nntpcache.c:208:Connection refused: SIGSEGV! > Jul 28 08:50:53 michelob nntpcache-client[16709]: spc-isp-van-uas-04-29.sprint.ca exit articles 0 groups 1 > Jul 28 08:50:53 michelob nntpcache-client[16709]: spc-isp-van-uas-04-29.sprint.ca times user 0.00 system 0.01 elapsed 1.00 > Jul 28 08:50:53 michelob nntpcached[16709]: clean shutdown with error 1. dumping core for debug analysis > > Case 2: > ------- > > Similar to above, but different message during the actual SIGSEGV. Note a "LISTGROUP" in > both cases.. > > Jul 28 08:52:15 michelob nntpcache-client[16723]: sockets.c:427: <- GROUP alt.binaries.pictures > .erotica.early-teens > Jul 28 08:52:15 michelob nntpcache-client[16723]: spc-isp-hfx-uas-03-29.sprint.ca group alt.bin > aries.pictures.erotic.children 1 > Jul 28 08:52:15 michelob nntpcache-client[16723]: sockets.c:427: <- LISTGROUP alt.binaries.pict > ures.erotica.early-teens > Jul 28 08:52:15 michelob nntpcache-client[16723]: sockets.c:153: server attach bypassed (server > tagged down) 57 seconds remaining before retry > Jul 28 08:52:15 michelob nntpcache-client[16723]: nntpcache.c:208:Bad file number: SIGSEGV! > Jul 28 08:52:15 michelob nntpcache-client[16723]: spc-isp-hfx-uas-03-29.sprint.ca exit articles > 1 groups 2 > Jul 28 08:52:15 michelob nntpcache-client[16723]: spc-isp-hfx-uas-03-29.sprint.ca times user 0. > 00 system 0.03 elapsed 18.00 > Jul 28 08:52:15 michelob nntpcached[16723]: clean shutdown with error 1. dumping core for debug > analysis > > > I've had about 8 of the "connectiong refused SIGSEGV" mesages (all with preceding LISTGROUP > commands) and 2 of the "Bad file number" segv's in the last 3 hours. > > > Case 3: > ------- > > Jul 28 10:45:22 michelob nntpcache-client[17609]: sockets.c:427: <- MM-$TJ(JFJR%!DO'YN&SCKP > > Whole bunch of weird crap like that... > followed by: > > Jul 28 10:45:22 michelob nntpcached[352]: nntpcache.c:400: starting client task > Jul 28 10:45:22 michelob nntpcached[352]: nntpcache.c:208:Interrupted system call: SIGSEGV! > > then immediately lots more of the weird stuff. Note that nntpcached[352] was the parent > process. At that point, it appears that the children finished doing there business, > but with some failures: > > Jul 28 10:45:23 michelob nntpcache-client[18561]: article.c:259: adding <35c4867f.256884@news.u > swest.net>:some-news-server.sprint.ca/alt/fan/oksana-bayul/small-tits/72460 to cache.history (failed) > > Jul 28 10:45:24 michelob nntpcache-client[17609]: sockets.c:427: <- M?T:ZU"0VBWERU^;6Y68#RW%LS1 > *.Q+LN`".M=T,.HRC[VYX>+Q7M> > > Lots more garbagte from that same child, but at this point I think we were dead in the > water, not accepting connections anymore. Most children seemed to finish sending articles > fine. Note that are about 25-40 entries of the garbage per second. Sometimes 300-400 > entries per second. > > Then: > > Jul 28 10:45:32 michelob nntpcache-client[17609]: sockets.c:463: => [some-news-server.sprint.ca] MKTWP)\, > /!*V&L>(XO$FK7?@W3(;5B#9J+_[5*Q46[+]PE<$EQQBO5O > > > Initial suspisions: One of the children went haywire and starting logging the contents of > its transfers to/from the backend content server. This is what the log files look like. > This is related to the parent process core dumping. > > The other suspision is that it was some sort of attack, but I don't think so.. > > I do see the contents of posts to bqackend sometimes: > > Jul 28 13:56:10 michelob nntpcache-client[24208]: sockets.c:463: => [some-news-server.sprint.ca] Message-ID: <35BE10A6.8D737998@jablowmi.com> > Jul 28 13:56:10 michelob nntpcache-client[24208]: sockets.c:463: => [some-news-server.sprint.ca] Date: Tue, 28 Jul 1998 13:55:50 -0400 > Jul 28 13:56:10 michelob nntpcache-client[24208]: sockets.c:463: => [some-news-server.sprint.ca] From: Heywood Jablowmi <heywood@jablowmi.com> > Jul 28 13:56:10 michelob nntpcache-client[24208]: sockets.c:463: => [some-news-server.sprint.ca] Reply-To: heywood@jablowmi.com > Jul 28 13:56:10 michelob nntpcache-client[24208]: sockets.c:463: => [some-news-server.sprint.ca] X-Mailer: Mozilla 4.05 [en] (WinNT; I) > Jul 28 13:56:10 michelob nntpcache-client[24208]: sockets.c:463: => [some-news-server.sprint.ca] MIME-Version: 1.0 > Jul 28 13:56:10 michelob nntpcache-client[24208]: sockets.c:463: => [some-news-server.sprint.ca] Newsgroups: alt.sex.bondage.female-admins-nntp > Jul 28 13:56:10 michelob nntpcache-client[24208]: sockets.c:463: => [some-news-server.sprint.ca] Subject: Horney Harbor $10 Hooker > Jul 28 13:56:10 michelob nntpcache-client[24208]: sockets.c:463: => [some-news-server.sprint.ca] Content-Type: text/plain; charset=us-ascii > Jul 28 13:56:10 michelob nntpcache-client[24208]: sockets.c:463: => [some-news-server.sprint.ca] Content-Transfer-Encoding: 7bit > Jul 28 13:56:10 michelob nntpcache-client[24208]: sockets.c:463: => [some-news-server.sprint.ca] Organization: Sprint Canada Inc. > Jul 28 13:56:10 michelob nntpcache-client[24208]: sockets.c:463: => [some-news-server.sprint.ca] Cache-Post-Path: michelob!unknown@spc-isp-tor-uas-15-13.sprint.ca > Jul 28 13:56:10 michelob nntpcache-client[24208]: sockets.c:463: => [some-news-server.sprint.ca] X-Cache: nntpcache 2.3.2.1 (see http://www.nntpcache.org/) > Jul 28 13:56:10 michelob nntpcache-client[24208]: sockets.c:463: => [some-news-server.sprint.ca] Horney Harbor $10 Hooker > Jul 28 13:56:10 michelob nntpcache-client[24208]: sockets.c:463: => [some-news-server.sprint.ca] http://www.horneyharbor.com > Jul 28 13:56:10 michelob nntpcache-client[24208]: sockets.c:463: => [some-news-server.sprint.ca] Really, $10 a month gets you everything the fancy ass sites give you and more. > Jul 28 13:56:10 michelob nntpcache-client[24208]: sockets.c:463: => [some-news-server.sprint.ca] That includes all of the following at no extra cost. > Jul 28 13:56:10 michelob nntpcache-client[24208]: sockets.c:463: => [some-news-server.sprint.ca] Free LIve Sex Rooms (hardcore amateur rooms) > Jul 28 13:56:10 michelob nntpcache-client[24208]: sockets.c:463: => [some-news-server.sprint.ca] 2800+ video channels (14 categories xxx uncensored) > Jul 28 13:56:10 michelob nntpcache-client[24208]: sockets.c:463: => [some-news-server.sprint.ca] huge XXX archive (37000 pics) > Jul 28 13:56:10 michelob nntpcache-client[24208]: sockets.c:463: => [some-news-server.sprint.ca] Personal Kinky ads (with reply function) > Jul 28 13:56:10 michelob nntpcache-client[24208]: sockets.c:463: => [some-news-server.sprint.ca] Free email box (full function) > Jul 28 13:56:10 michelob nntpcache-client[24208]: sockets.c:463: => [some-news-server.sprint.ca] Free Chat (check it out 17 horney rooms plus private rooms) > Jul 28 13:56:10 michelob nntpcache-client[24208]: sockets.c:463: => [some-news-server.sprint.ca] Three days free trial. > Jul 28 13:56:10 michelob nntpcache-client[24208]: sockets.c:463: => [some-news-server.sprint.ca] follow this link=> http://www.horneyharbor.com > Jul 28 13:56:10 michelob nntpcache-client[24208]: sockets.c:463: => [some-news-server.sprint.ca] How do we do it. Easy we don't pay $20 advertising per member to find you like the fancy ass boy > > Nice way to catch spammers, but otherwise creates massive log files. > This doesn't seem to be the norm in the logging, since it usually just > says "ARTICLE -> .." > > So, is this normal?? > > I also get TONS of SEGV's in the nntpcache-nocem portion. I had to turn > it off. That doesn't really bother me though, since we'll be getting rid > of those spams upstream in the real-near-term. :-) > > Many thanks to anyone that can provide some insight/advice. (And a > beer if you live in the Toronto area! ;-)) > > -- > ------------------------------------------------------------------------ > Mark Mayo mark@vmunix.com > RingZero Comp. http://www.vmunix.com/mark > ------------------------------------------------------------------------ > "The problem is how do you build tools that understand your programs > at a deeper semantic level." - James Gosling > >
