I've actually been running the most recent master 
(0fd248c9c305e7b1bca086be1e883b5d47430e99) on these instances.

I agree that CPU steal on t1.micros is a likely culprit. Fortunately I'm 
not using such a handicapped instance in production; it was just a 
throwaway test server. :)


Patrick.

On Friday, February 21, 2014 4:17:43 PM UTC-8, Daniel Farina wrote:
>
> On Fri, Feb 21, 2014 at 3:46 PM,  <[email protected] <javascript:>> 
> wrote: 
> > Just to add something to the mix, I noticed this same issue when running 
> > WAL-E on a t1.micro instance backing up to S3 in the same region - it 
> looks 
> > like the network performance is just too low. 
> > 
> > When running the same setup on an m3.large instance it worked just fine, 
> > first try. 
>
> I also am suspicious about the extensive CPU used by WAL-E.  t1.micros 
> are pretty tiny so it's hard for me to guess, but please do try 
> installing the latest master -- I've *just* committed some very large 
> performance optimizations -- and reporting back if you can. 
>

-- 
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/groups/opt_out.

Reply via email to