> My terrible options are as follows: Another option is to run LVM+[XJ]FS on, say, RAID5. Then if a disk is failing, you can simply pull it and run in degraded mode until the replacement arrives. Of course that doesn't help you *now*...
BTW, the software raid tools for linux actually support growing, even in a RAID5 configuration. Now before you get excited, md can not currently change the number of devices in a RAID5, though that capability is apparently being worked on (this would allow you to grow the RAID5 device by adding drives). Rather, if you replace the current drives, one at a time, with larger ones, you can then grow the RAID5 set to fill the new space. So, md isn't as restrictive as some people think, though arguably still not as flexible as a pure LVM volume group. Still, it would be nice if JFS could shrink. I use to run ext3 for everything simply for that reason. Well, and I have a bad history with XFS and ReiserFS (yes it was from a long time ago, but it was a production system and I don't forgive easily) I've never been bit by ext3. But the deletes really were just to slow, especially for HD sized shows. Now I run everything on JFS/LVM/RAID5. No problems yet. Oddly, I think JFS on AIX can shrink, so maybe it is a feature that will be implemented some day on Linux. (Maybe I'm remembering that wrong) -poul
_______________________________________________ mythtv-users mailing list mythtv-users@mythtv.org http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-users