Op dinsdag 27-10-2009 om 19:39 uur [tijdzone -0500], schreef Leslie Rhorer: > > Just rsync the two boxes already, with or without ssh. That's assuming > > the data is mostly static (videos). If you have other data as well you > > might want to use rdiff-backup for that, as it can keep multiple > > versions. > > > > Get some kind of backplane for the disks so you can swap them out > > without powering down the box. (Not because you can't live with 10min > > downtime when a disk fails but because you don't want to spin down the > > *other* disks.) > > > > Bonus: If the primary server fails but its disks are fine you can just > > stick all of them in the backup server temporarily, boot and be up and > > running almost immediately. > > Well, depending on the system and exactly what dies, it may be > simpler just to fix or replace the server. The big bonus comes when the > primary array fails, though. If the OS files and directories (/boot, /etc, > /var, /home, etc.) are all on a separate drive system and the primary array > contains only application data, then upon total failure of the primary > array, one may simply umount the main array and use NFS to temporarily mount > the remote file system wherever the primary array used to be. The entire > production system is then up and running as usual, while the sysadmin > figures out what the heck went wrong with the primary array. When the array > is fixed, an rsync (hopefully a brief one) can copy all the new / changed > files to the main array. Then umount the NFS share and mount the main array > again. > Agreed. Just a pity the OP thinks NFS is useless old bloadware.... Rudy -- To unsubscribe from this list: send the line "unsubscribe linux-raid" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
