NetBSD-Bugs archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: kern/37955: Kernel Panic NetBSD 4.0 blkfree

The following reply was made to PR kern/37955; it has been noted by GNATS.

From: David Holland <>
To: "V.L. van den Berg" <>
Subject: Re: kern/37955: Kernel Panic NetBSD 4.0 blkfree
Date: Sun, 4 Jan 2009 01:04:26 +0000

 On Tue, Mar 11, 2008 at 11:11:43PM +0100, V.L. van den Berg wrote:
 > I'll check if there is enough space left on the device to perform the  
 > operations you mentioned. Can you give me instructions on how to copy  
 > the fs contents into one file? I've never done this before.
 Hmm. Apparently I never answered this.
    # dd if=/dev/rsd0a of=somefile.img bs=1048756
 (or whichever partition if it's not sd0a) copies the filesystem image
 to the file somefile.img.
 Then you should be able to run fsck on somefile.img. But you need
 enough free space to hold the entire disk, including the unused bits.
 Anyway, PR 35790 is now confirmed fixed, so if your problem was/is
 related it should also be fixed, and the fix should at least be in
 5.0_BETA if not also in 4.0_STABLE.
 (previous context)
 > David Holland schreef:
 >> On Mon, Mar 10, 2008 at 01:31:52PM +0100, Vincent van den Berg wrote:
 >>  > >Just to check, when you say clean, you mean like this, right?
 >>  > >[snip]
 >>  >
 >>  > Hi David,
 >>  >  > Yes, I forced fsck on all volumes. It goes through them like you  
 >>  > described (Phases etc.) and it gives no errors.
 >> Sigh. It would have been much simpler if the problem turned to be just
 >> that the fs was corrupt. Or at least, simpler to make it work again.
 >> I'm wondering if this is the same as or related to kern/37590.
 >> Do you have enough disk space to make a copy of the contents of the fs
 >> in a file? If you do, do that, fsck it, then use vnd(4) to mount the
 >> file as a fs, and try running rsnapshot or rsync or whatnot. While
 >> that's useless for backing up, if it works reliably, then we can be
 >> pretty sure it's a usb issue and not a filesystem issue, which rules
 >> out lots and lots of code that might be causing problems. On the other
 >> hand, if it blows up that way too, then we can be pretty sure it's
 >> *not* a usb issue.
 >> Also, when sending mail, cc:; then it'll go 
 >> into
 >> the PR database so other developers can see, which is a good thing
 >> anytime something mysterious is happening.
 David A. Holland

Home | Main Index | Thread Index | Old Index