[SGVLUG] Resonable way to copy some files?

[Michael Proctor-Smith]

So I finally had one of my drives(in an LVM appended two drive array a
120GB and 160GB) drive in my main tivo box crap out on me. Luckily it
seemed to be a problem that made the drive unwriteable but readable in
fact the system was online at the time(well it is never offline
otherwise how could it record tv) and I notice some problems while I
was watching one of the remote frontends. So I logged in and saw a
high load and most of the cpu time being spent waiting for IO. I
checked the dmesg and was greeted with a string of /dev/hdb related
failure and ext3 block inaccessible errors.

Well I had about 148GB of TV shows (little less then ~70 hours), that
some of which I really wanted. So I gave coping the files off the
system a shot before I rebooted the machine hoping the drive would
recover on reboot(It happens). Well I got all of it off without any
errors(I had my HD mythbackend box already setup with nfs share so
148GB at about 10MB/s later I reboot the machine.

No joy in michaelvile, the drive is not writeable. So I now have
3hours till Thursday night shows start and I want to recorded like
like 8-10 shows over the course of the night.

So I happen to have a 3ware 7850 card(8 port hardware raid) and the
five iffy (160GB) drives that had been part of an array I replaced and
the owner did not want. So quickly I rip apart the box and throw it
back together. I figure the drives are iffy I will configure it as a 3
drive raid 5 array and then have two hot spares .

Anyway enough background I get the system back online the array and
lvm(for the future) configured and the mkfs run in time(well almost I
missed one show at 7 but it was on cable and aired again later in the
night). Anyway I knew that the array was still rebuilding and therefor
would only have limited IO (hdparm at the time said 12MB/s) so I
waited till after it was done recording to start copying files back
from the other machine (via nfs). Well about 2:00AM as I was going to
bed I started it copying. Well I woke up in the morning to an unsinked
interrupt handler kernel panic. The system array was not fully
initized until some time this afternoon.

So I had 30GB of the 148GB back and not being a scripting genus I
wanted the easiest/fastest way to copy back only the files that had
not been copied before. rsync? anyway that is what I ended up using.

Anyway the load average on the machine around 8-11 again mostly
waiting for IO. The question is what would the best solution have been
and is 100+ interrupts per second(from the 3ware card) a lot?

Sorry I wrote my life story.