On Thu, Sep 15, 2005 at 08:55:13PM +0100, David wrote: > Sorry, I have to disagree - and pretty strongly too!! > > Resizing a filesystem with a known bad disk is *begging* for a total > toast situation (to use an expression I've grown rather fond of!). > You're basically asking the system to hammer a failing disk with read > and re-reads!
I've taken a different approach that has worked very well for my desires. For me, I don't have anything on one system or medium that I don't want to lose. raid5 still can go bad, be it a power surge that kills all disk or an accidental delete. I have a failing disk, not meaning it's failed in any way yet. I run daily checks with smartctl and I am notified if a disc ever starts to reallocate blocks. The drive in question is starting to reallocate blocks, which means it hasn't failed, isn't failling, but will fail in days to months. What I had planned on doing is when I noticed a drive starting to do this I'd just remove it from LVM, except that's not possible since I can't shrink my filesystem, and thus the reason for the start of this thread -- To warn people that you can't just free enough disk space and pvremove a disk if you're running XFS, JFS, or Reiser4. Now if I had been using ReiserFS then I could have resized the filesystem and done a pvremove, and tossed the disk away before it even lost a single byte. --Brandon
_______________________________________________ mythtv-users mailing list [email protected] http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-users
