Hi Richard, Richard Elling wrote: > Occasionally the topic arises about what to do when a file is > corrupted. ZFS will tell you about it, but what then? Usually > the conversation then degenerates into how some people can > tolerate broken mp3 files or whatever. > > Well, the other day I found a corrupted file which gave me an > opportunity to test a little hypothesis on how to recover what > you can recover. Details are in my blog: > http://blogs.sun.com/relling/entry/holy_smokes_a_holey_file > > There is an opportunity here, for someone with some spare time, > to come up with a more clever solution than my dd script. hint... > hint... > -- richard > > Would it help if you had the block number (i.e., disk location) of the block that is corrupted? zdb might tell you this. I have a way to do it, I think, but don't want to test it because I don't want to corrupt a file on purpose. I am writing a paper (actually, done with first draft) that allows you to find the data for a given file on the raw disk (i.e., file system is not mounted). I plan on presenting this at osdevcon in Prague in June. I am looking for reviewers. If you are interested, please send me email and I'll send you a copy. The method I use is quite a bit more complex than using dd. It involves using a modified zdb and modified mdb together. I think it would work for this type of problem. (Then again, if zfs completely wipes out the corrupted block, it won't help). If I have time, I'll try corrupting a few bits in a file and see if my method works to get the corrupted block.
thanks, max > _______________________________________________ > zfs-discuss mailing list > zfs-discuss@opensolaris.org > http://mail.opensolaris.org/mailman/listinfo/zfs-discuss > > _______________________________________________ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
