Xref: utzoo unix-pc.general:914 unix-pc.uucp:30 Path: utzoo!utgpu!water!watmath!uunet!zorch!pacbell!maxepr!ken From: ken@maxepr.UUCP (Ken Brassler) Newsgroups: unix-pc.general,unix-pc.uucp Subject: Re: 3.51a keeps dying; even HDB is a mess; any ideas out there? Message-ID: <513@maxepr.UUCP> Date: 25 Jun 88 20:00:15 GMT References: <422@kosman.UUCP> Reply-To: ken@maxepr.UUCP (Ken Brassler) Organization: Brassler Engineering Co., Mill Valley, CA Lines: 33 In article <422@kosman.UUCP> kevin@kosman.UUCP (Kevin O'Gorman) writes: >I have been having bad results lately, and don't quite know what to blame >it on. >I am distressed because my machine keeps freezing on me. It seems pretty >clearly related to uucp traffic. >It was just at the end of an >incoming uucp call (my unix-pc feed, to be exact), and it left something >for cunbatch in /lost+found. My machine has run continuously and flawlessly for 2 1/2 years. A few weeks ago a had my first kernel panic, due to a kernel parity error. I used the reset button to recover. Later that same day, my machine crashed (froze, locked-up, etc.) twice while receiving news. The second time, I was using 'rn' at the same time, and after reset, I found that my .newsrc file now contained the text of a news article. Since both crashes occurred while uucico was running, I surmised that the disk image of uucico had probably been partially overwritten with garbage during the first kernel panic. I reloaded new copies of all the uucp executables, uucico, uux, uuxqt, and uucp from my archives, and the problem immediately disappeared. Personally, I think that kernel crashes are due to a dram address, where the kernel is loaded, missing a refresh cycle, or being hit by a cosmic ray (not completely a joke). If no damage was done to the files on the hard disk during the crash, you can get by with a reset. If crashes increase in frequency, I think it's time to reformat and reload the hard disk. -- Ken Brassler {ihnp4|qantel|pyramid|lll-crg}!pacbell!maxepr!ken