Return to BSD News archive
Newsgroups: comp.unix.bsd Path: sserve!manuel.anu.edu.au!munnari.oz.au!sgiblab!spool.mu.edu!agate!usenet.ins.cwru.edu!gatech!concert!sas!mozart.unx.sas.com!torpid.unx.sas.com!sastdr From: sastdr@torpid.unx.sas.com (Thomas David Rivers) Subject: Re: Corrupted directory Sender: news@unx.sas.com (Noter of Newsworthy Events) Message-ID: <Bw62ps.n57@unx.sas.com> Date: Thu, 15 Oct 1992 14:25:04 GMT References: <1bd165INNrfe@iskut.ucs.ubc.ca> <Bw4Fv7.M4L@unx.sas.com> <1bi6soINN8au@iskut.ucs.ubc.ca> Nntp-Posting-Host: torpid.unx.sas.com Organization: SAS Institute Inc. Lines: 58 In article <1bi6soINN8au@iskut.ucs.ubc.ca> felawka@sitka.triumf.ca (Larry Felawka) writes: >In article <Bw4Fv7.M4L@unx.sas.com>, sastdr@torpid.unx.sas.com (Thomas David Rivers) writes: >|> >|> Regarding this infinite reboot cycle; I recently applied the entire >|> beta patchkit (from a new, clean, installation of 386bsd 0.1) and >|> began getting this problem. >|> >|> I.e. If I have to shut down for some reason, or the power goes out; the >|> machine reboots. Root gets fsck'd and repaired; and the >|> machine reboots again. Then there is another problem with the >|> fsck of root, and the machine reboots, etc... encountering the >|> same problem over and over. >|> >|> I've now seen this happen on two machines to which I applied the beta >|> patchkit. It did *not* happen on these machines using the kernel I >|> had patched together from news articles. This means (at least to me) >|> there is some patch applied by the patchkit which I had not otherwise >|> applied in my kernel, that is causing the problem. >|> >|> I've seen it on a SCSI disk and IDE disk, so I don't think it's in the >|> driver... could it be some of the changes to init.c? >|> > >Only indirectly ... the problem comes when you do the "make install" for >"init". It appears that at this point the file system has become corrupted >and if you reboot you get the behaviour you have described. One can recover >by inserting the fixit disk, rebooting from the floppy and doing >"fsck /dev/rwd0a". The clean way to do the patch installation for init (and >other programs that are running while you are patching them) is probably to >do the "make", shut down and do the installation using the fixit disk. > >-- >Larry Felawka While that may be the case; I have seen it occur several days (and *many* reboots) after I changed init. There must be something else going on as well. The problem is not consistent; just every now and then, when rebooting after the machine locks up (leaving the file system in a bad state) you can fall into this infinite-boot-loop. Also, I noticed that, with the new init, when the fsck of root fails, it drops you into single user mode, but the rc script is still running, as evidenced by a message "starting network" appearing over my single-user shell prompt. Then; it reboots... Could this be causing the problem (i.e. we get "part of the way" up to multi-user, and munge the file system again before the reboot succeeds.) - Dave Rivers - (rivers@ponds.uucp (home)) (sastdr@unx.sas.com (work)) -- UPDATE ALL INFORMATION AND POD INTO COSMOS - Federal Express