DRBD halt copying file internally

Hey all - Just setting up my first set of HA machines. Things are working well, except for one snag with DRBD. If I try to copy a set of largeish files from one location of a drbd volume to another location on the same volume, I get an io error and a halt.

This is pretty repeatable. So like:

cp /srv/www/mywebsite/podcast/*mp3 /srv/www/testwebsite/podcast/

Will result in the following (as seen from lish):

linode-secondary login: end_request: I/O error, dev xvdd, sector 27082744

end_request: I/O error, dev xvdd, sector 27082744

block drbd1: drbdmdsyncpageio(,27082744s,WRITE) failed!

block drbd1: meta data update failed!

block drbd1: Local IO failed in drbdmdsync. Detaching…

block drbd1: Sending state for detaching disk failed

end_request: I/O error, dev xvdc, sector 2047992

end_request: I/O error, dev xvdc, sector 2047992

block drbd0: drbdmdsyncpageio(,2047992s,WRITE) failed!

block drbd0: meta data update failed!

block drbd0: Local IO failed in drbdmdsync. Detaching…

block drbd0: Sending state for detaching disk failed

block drbd0: PingAck did not arrive in time.

block drbd0: PingAck did not arrive in time.

block drbd0: PingAck did not arrive in time.

block drbd1: PingAck did not arrive in time.

block drbd0: PingAck did not arrive in time.

block drbd1: PingAck did not arrive in time.

block drbd0: PingAck did not arrive in time.

block drbd0: PingAck did not arrive in time.

block drbd0: PingAck did not arrive in time.

block drbd1: PingAck did not arrive in time.

block drbd0: Sending state for detaching disk failed

block drbd1: Sending state for detaching disk failed

block drbd0: PingAck did not arrive in time.

block drbd1: PingAck did not arrive in time.

block drbd1: error receiving Data, l: 4120!

block drbd0: PingAck did not arrive in time.

Anybody run across this? Thoughts? It's ubuntu 10.04LTS 64bit, 3.0.4-linode37