pausv@sssab.se (Paul Svensson) (03/18/91)
I'm running C News 1-Dec-1990 on an RM Nimbus with Interactive 386/ix 2.0.2. It's works fairly well, but with some annoying problems I can't understand. Sometimes newsrun dies, I get a mail from cron saying "newsrun: Alarm call", and a LOCKTM* file is left littering $NEWSCTL. Otherwise everything seems ok; nothing in errlog, nothing in in.coming/bad, no processes hanging around waiting for a a lock. Removing the litter keeps checknews from complaining, and that's all I've been able to do about it. The system has been running like this a few weeks, and this has been happening about every other day or so. Today it got worse. I got about a hundred mail from cron, saying that newsrun got an alarm, and as many from news, saying that relaynews exited with status 141 (ST_SHORT|ST_ACCESS|ST_DROPPED|ST_NEEDATTN|1). in.coming/bad was overflowing, and $NEWSCTL was full of LOCKTM* files. errlog contained the single line "sh: 20632 bus error - core dumped"; I found a core file with the same date as errlog on $NEWSARTS, it looks like smail. ps showed me two sendbatches running; one of them was sleeping for the lock, the other one didn't seem to be doing anything. There's no shortage of neither disk space nor inodes anywhere on the system. My reflex reaction was to get news running again, I could always figure out what was wrong later, so I just cleaned up the mess, removed the locks, killed the processes, put a few bad batches back up and run newsrun, expecting to see something fail. It went through without a squeak; in fact, all the batches worked just fine when I reran them. I'm trying to figure out from the source what could cause relaynews to give exit status 141, but I'm totally at loss. I don't understand where the 1 is coming from, I don't see how ST_NEEDATTN or ST_SHORT could be set by any problem that could go away by it self, and I can't at all figure out how relay/procart.c works, only that it sets ST_ACCESS when it can't unlink some file. And who is sending a SIGALRM to newsrun ??? Could someone please enlighten me, or at least give me a clue where to look if this happens again ? /Paul -- Paul Svensson _ /| - Every absurdity needs a champion to defend it - SM5SJS \'o.0' Scandinavian System Support Fax: +46 13 115193 paul@sssab.se =(___)= Box 535 _ Phone: +46 13 111660 sunic!sssab!paul U SE-581 06 Linkoping, Sweden Home: +46 13 121021