On Nov 11,  3:29pm, Boris Boesler wrote:
> Subject: Probs with ffs_update
> It happend again:
>
> ffs_update: bad indirect add(1): inode=98124 4194304/00400000
> ip=f14dd500 adr=f14dd5b0
>
> inode vp: type VBLK, usecount 664, writecnt 0, refcount 27.
> tag VT_UFS: no 98 on dev 24 , 0
> ffs_update: bad indirect addr patched inode=98124
>
> Last time this happened the bootblock was gone with the wind
> and I had to reinstall BSD. I don't want to do this again !!
>
> The RPC's hardware:
> 20+2MB, connect32+cfp1080s, kernel #4626, 1.2-beta sets
>
>
> Any comments?

It does that. We wish we knew why. Newish kernels attempt to fix the problem by
zeroing the indirect pointer. This may end up trashing really long files which
actually have it set to something non-zero in the first place. Normally you get
away with it and just get a lot of junk in your log files.

Cheers

Rob

--