Mailing-List: contact dev-help@couchdb.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@couchdb.apache.org
Received-SPF: pass (athena.apache.org: domain of randall.leeds@gmail.com
 designates 209.85.215.52 as permitted sender)
MIME-Version: 1.0
In-Reply-To: <1333665006580-7441597.post@n2.nabble.com>
References: <1333665006580-7441597.post@n2.nabble.com>
Date: Thu, 5 Apr 2012 23:56:42 -0700
Message-ID: 
 <CAAL6JQj-i15X66n15WQ-5X0VQ3U_TpwFE03Fya+7yewnZHX5LA@mail.gmail.com>
Subject: Re: 2 Posts at once causes problems
From: Randall Leeds <randall.leeds@gmail.com>
To: dev@couchdb.apache.org
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

On Thu, Apr 5, 2012 at 15:30, ashgavs <ashgavs@gmail.com> wrote:
> I did some testing on my client, and made two posts that get posted at
> exactly the same time. =C2=A0We have long polling enabled so I waited for=
 a GET
> to see if both made it successfully into the DB. =C2=A0Both posts went ou=
t, but
> we only got one GET back. So I checked the database, it had both document=
s.
> So what happened? Why did one get lost?

If you re-issue the long poll GET with the last_seq returned do you
see the other POST?
It is expected that long polling may only give you the first of the
two POST if the GET was initiated before the POST.

>
> The robot makes literally thousands of posts in a short period of time. I=
ts
> constantly sending the DB information. My theory is that if a user and th=
e
> robot make a post at exactly the same time, its possible that the user wi=
ll
> not GET back both posts from the DB, almost like a threading issue. Do yo=
u
> have any thoughts? Have you seen anything like this before?

If you are starting the GET at the same time as the POST there could
have been a small race condition:
https://issues.apache.org/jira/browse/COUCHDB-1363

-Randall