Sudden OAuth failures from a specific IP address

0 views
Skip to first unread message

timwhitlock

unread,
Nov 25, 2009, 12:11:43 PM11/25/09
to Twitter Development Talk
Has anyone found their existing code suddenly getting OAuth failures?
I'm getting a 401 ("Failed to validate oauth signature and token")

Normally I'd assume that I've been an idiot and screwed up my code -
Except that the code is unchanged from yesterday when it was working
perfectly. I've also had the same code in production for months with
no such errors, and those sites are still operating fine.

This points to a problem with my IP.

Can anyone at Twitter help? Have I been blocked?
It's only the OAuth call that is failing. Other non-authed requests
are working ok.
Message has been deleted

Ryan Sarver

unread,
Nov 25, 2009, 11:45:29 PM11/25/09
to twitter-deve...@googlegroups.com
Tim,

can you provide us with your IP so we can look into it?

Thanks, Ryan

On Wednesday, November 25, 2009, timwhitlock
Message has been deleted

timwhitlock

unread,
Nov 28, 2009, 8:10:43 AM11/28/09
to Twitter Development Talk
Hey Ryan
I think I've found the problem. Our app was making two calls to get a
request token in very quick succession which resulted in the same
nonce.
I think this explains the 401, so not an IP problem at all.
Sorry if I wasted your time.





On Nov 26, 10:16 am, timwhitlock <tim.whitl...@publicreative.com>
wrote:
> Thanks Ryan.
> The IP that results in failures is 87.224.88.186
> I've run the exact same code on three other servers and the response
> is fine.
>
> We recently mucked about with some DNS settings, so perhaps you run
> some kind of reverse lookup, and something looks suspicious?
> Let me know if there's anything I can sort out there. The IP is
> static, but it's not a hosting provider. This is our office
> development server's public facing address.
>
> Tim.

timwhitlock

unread,
Dec 3, 2009, 8:12:19 AM12/3/09
to Twitter Development Talk
Turns out after all that - it was my system clock being wrong.
Our dev server clock keeps screwing up. The 401 wasn't descriptive
enough for me to diagnose this.
It was actually Amazon S3's descriptive error that let me to this
solution
Reply all
Reply to author
Forward
0 new messages