Date: Wed, 23 Jun 2004 09:42:05 -0400 (EDT) From: Jerry McAllister <jerrymc@clunix.cl.msu.edu> To: ruben@bloemgarten.demon.nl Cc: questions@freebsd.org Subject: Re: md5 of a filesystem / verifying filesystem integrity after Message-ID: <200406231342.i5NDg6g18473@clunix.cl.msu.edu> In-Reply-To: <E1Bd7cM-000NAA-00@post-21.mail.nl.demon.net> from "Ruben Bloemgarten" at Jun 23, 2004 03:14:57 PM
next in thread | previous in thread | raw e-mail | index | archive | help
> > Hi all, > > Does someone know how to reliably run a checksum of sorts on a filesystem, > to be able > > to verify filesystem integrity after a restore from dump level 0 has > occurred? Unless you made a checksum of everything before doing the dump and made absolutely no changes, there is no way that I know of. Even some little things such as block arrangements being a little different after a restore - irrelevant to the integrity of the files - would make a checksum come out differently if the whole filesystem were checksummed. You can do an fsck(8) and see if there are any problems in file pointers. But, that is hardly worth bothering with since it is done at boot time anyway and it doesn't check the content of the files, only the pointers/chains. On some versions of dump either older or from other sources (vendors) dump had a verify flag that would read back everything and check it with what is on disk. I don't see it in FreeBSD's dump. Anyway, it was almost useless since the system would have to be down for the whole dump and the whole verify pass. Most people do dumps with the system up and running and in this case the files can change during the dump and verify times, thus making verify always fail. It is a rare system that can afford to be down long enough to do even a single user mode dump, let alone adding on the verify. So, it doesn't surprise me to see that option gone. ////jerry > > Thanks, > > Ruben >
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200406231342.i5NDg6g18473>