On Sun, Dec 7, 2008 at 1:37 PM, Matt Burkhardt <mlb AT imparisystems DOT com> 
wrote:
>    localhost /samba/bigdrive/Music lev 0: partial taper:  While writing data
> block to S3: Too many retries; last message was 'Your socket connection to
> the server was not read from or written to within the timeout period. Idle
> connections will be closed.' (RequestTimeout) (HTTP 400) (after 14 retries)

Hmm, I recall having this kind of trouble way back when S3 was still
in beta, but haven't seen it since.  I assumed it was due to a bug on
the Amazon side.

The S3 device does not begin sending data until it has a full block
available, and then sends that block as quickly as the network will
permit, so this sort of timeout cannot be caused by slowness on the
client or anything like that.

Have you modified your blocksize from the default 10M?  Is there some
reason that a TCP connection to Amazon would stall out for more than
60 seconds?  Can you upload multi-megabyte files to S3 using other
utilities (e.g., JetS3t)?

Dustin

-- 
Storage Software Engineer
http://www.zmanda.com