Tino Schwarze wrote:
>
>> A missing member doesn't bother raid1 at all.
>
> It bothered mdadm --monitor! It always complained, there seemed to be no
> way to fix it, so monitoring became useless.
You can just peek at /proc/mdstat once in a while.
>> The way I do it is to create a raid1 with 3 members, one specified as
>> 'missing'. Then once a week I put in a disk and 'mdadm --add ...' the
>> partition to the array. When it finished mirroring (a little over 2
>> hours for a 750 gig SATA), I momentarily stop backuppc and unmount the
>> array so the filesystem is clean, then fail and remove the partition and
>> rotate that disk offsite. That way I always have a pair of mirrored
>> disks locally and one or two slightly out of data offsite.
>
> Bad luck. My backuppc file system is already at 900 GB and located in an
> LVM volume which is on a HW RAID5 of 3x750 GB SATA disks.
You can get 1 TB SATA drives these days - and trayless hot-swap enclosures.
> I wonder
> whether something similar would be possible with LVM snapshots or
> somesuch... but I suppose, I would pay yet another performance penalty.
Theoretically yes, but I've never seen enough documentation on
recovering a corrupted LVM to trust my ability to do it. If I ever
have time to build something that has zfs (solaris, freebsd, or maybe if
I wait long enough, OSX), I'd like to try its snapshot send/receive
facility.
--
Les Mikesell
[EMAIL PROTECTED]
-------------------------------------------------------------------------
This SF.net email is sponsored by the 2008 JavaOne(SM) Conference
Don't miss this year's exciting event. There's still time to save $100.
Use priority code J8TL2D2.
http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone
_______________________________________________
BackupPC-users mailing list
BackupPC-users@lists.sourceforge.net
List: https://lists.sourceforge.net/lists/listinfo/backuppc-users
Wiki: http://backuppc.wiki.sourceforge.net
Project: http://backuppc.sourceforge.net/
