No, my problem is entirely with AWS machines that are in the same region as the S3 bucket. I have been running WAL-E from us-west-2 (Oregon) machines on an Oregon S3 bucket and failing until I switch to US Standard.
My connection between WAL-E and its S3 bucket is about as good as it gets. It's not the network reliability; something else is wrong. On Fri, Mar 14, 2014 at 1:23 PM, Daniel Farina <[email protected]> wrote: > On Fri, Mar 14, 2014 at 1:19 PM, Dan Fairs <[email protected]> wrote: > >>> For what it's worth, we're doing everything using a bucket in eu-west-1 > >>> (Ireland). > >> > >> Left unstated but assumed by the context, you are also using instances > >> in eu-west-1 right? > > > > > > Actually, no - we have physical hardware, we're not using EC2 at all. > Our machines are colocated in the UK. > > Okay, this may put you in the same boat as David Weaver, in that you > may be getting into S3 from outside AWS and that impacts reliability > of the connection. > > I think cleaning this up by suppressing the backtrace and adding a > HINT may be the right way to get the information to the relevant > people at the right time. Something like: > > HINT: Connection Reset by Peer is more frequent when using S3 from > outside the corresponding AWS region, but the upload will be re-tried. > > Not too happy with the wordsmithing, but that's the gist. Suggestions > welcome. Also, a patch. > -- You received this message because you are subscribed to the Google Groups "wal-e" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.
