Mailing-List: contact user-help@flink.apache.org; run by ezmlm
Precedence: bulk
Reply-To: user@flink.apache.org
MIME-Version: 1.0
In-Reply-To: <897438718.52960.1485436243035.JavaMail.zimbra@nododos.com>
References: <897438718.52960.1485436243035.JavaMail.zimbra@nododos.com>
From: Fabian Hueske <fhueske@gmail.com>
Date: Thu, 26 Jan 2017 15:21:45 +0100
Message-ID: <CAAdrtT0f5L=uTmQCKS-iK9M5CaR+FLQK5DhnMw77sxyqy4FYFQ@mail.gmail.com>
Subject: Re: State Descriptors / Queryable State Question
To: user@flink.apache.org
Content-Type: multipart/alternative; boundary=001a11468aa880a9b50547001348
archived-at: Thu, 26 Jan 2017 14:22:21 -0000

--001a11468aa880a9b50547001348
Content-Type: text/plain; charset=UTF-8

Hi Joe,

working on a KeyedStream means that the records are partitioned by that
key, i.e., all records with the same key are processed by the same thread.
Therefore, only on thread accesses the state for a particular key.
Other tasks do not have read or write access to the state of other tasks.

Best, Fabian

2017-01-26 14:10 GMT+01:00 Joe Olson <technology@nododos.com>:

> If I have a keyed stream going in to a N node Flink stream processor, and
> I had a job that was keeping a count using a ValueStateDescriptor (per
> key), would that descriptor be synchronized among all the nodes?
>
> i.e. Are the state descriptors interfaces (ValueStateDescriptor,
> ListStateDescriptor) threadsafe?If I expose that descriptor via the
> RunTimeContext, will I get a consistent value back from each of the nodes?
>
> Thanks!
>
>

--001a11468aa880a9b50547001348
Content-Type: text/html; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr"><div><div><div>Hi Joe,<br><br></div>working on a KeyedStre=
am means that the records are partitioned by that key, i.e., all records wi=
th the same key are processed by the same thread. <br>Therefore, only on th=
read accesses the state for a particular key.<br></div>Other tasks do not h=
ave read or write access to the state of other tasks.<br><br></div>Best, Fa=
bian<br></div><div class=3D"gmail_extra"><br><div class=3D"gmail_quote">201=
7-01-26 14:10 GMT+01:00 Joe Olson <span dir=3D"ltr">&lt;<a href=3D"mailto:t=
echnology@nododos.com" target=3D"_blank">technology@nododos.com</a>&gt;</sp=
an>:<br><blockquote class=3D"gmail_quote" style=3D"margin:0 0 0 .8ex;border=
-left:1px #ccc solid;padding-left:1ex"><div><div style=3D"font-family:arial=
,helvetica,sans-serif;font-size:12pt;color:#000000"><div>If I have a keyed =
stream going in to a N node Flink stream processor, and I had a job that wa=
s keeping a count using a ValueStateDescriptor (per key), would that descri=
ptor be synchronized among all the nodes?<br></div><div><br></div><div>i.e.=
 Are the state descriptors interfaces (ValueStateDescriptor, ListStateDescr=
iptor) threadsafe?If I expose that descriptor via the RunTimeContext, will =
I get a consistent value back from each of the nodes?<br></div><div><br></d=
iv><div>Thanks!<br></div><div><br></div></div></div></blockquote></div><br>=
</div>

--001a11468aa880a9b50547001348--