Return-Path: X-Original-To: archive-asf-public-internal@cust-asf2.ponee.io Delivered-To: archive-asf-public-internal@cust-asf2.ponee.io Received: from cust-asf.ponee.io (cust-asf.ponee.io [163.172.22.183]) by cust-asf2.ponee.io (Postfix) with ESMTP id 5A82B200D4C for ; Thu, 30 Nov 2017 10:48:10 +0100 (CET) Received: by cust-asf.ponee.io (Postfix) id 59030160BF4; Thu, 30 Nov 2017 09:48:10 +0000 (UTC) Delivered-To: archive-asf-public@cust-asf.ponee.io Received: from mail.apache.org (hermes.apache.org [140.211.11.3]) by cust-asf.ponee.io (Postfix) with SMTP id A0053160BEA for ; Thu, 30 Nov 2017 10:48:09 +0100 (CET) Received: (qmail 45890 invoked by uid 500); 30 Nov 2017 09:48:08 -0000 Mailing-List: contact user-help@flink.apache.org; run by ezmlm Precedence: bulk List-Help: List-Unsubscribe: List-Post: List-Id: Delivered-To: mailing list user@flink.apache.org Received: (qmail 45880 invoked by uid 99); 30 Nov 2017 09:48:08 -0000 Received: from mail-relay.apache.org (HELO mail-relay.apache.org) (140.211.11.15) by apache.org (qpsmtpd/0.29) with ESMTP; Thu, 30 Nov 2017 09:48:08 +0000 Received: from aljoschas-mbp.fritz.box (ip-2-205-80-95.web.vodafone.de [2.205.80.95]) by mail-relay.apache.org (ASF Mail Server at mail-relay.apache.org) with ESMTPSA id 296121A0044; Thu, 30 Nov 2017 09:48:06 +0000 (UTC) From: Aljoscha Krettek Message-Id: Content-Type: multipart/alternative; boundary="Apple-Mail=_5901C47F-0570-44FE-873A-010BBC902026" Mime-Version: 1.0 (Mac OS X Mail 11.1 \(3445.4.7\)) Subject: Re: Issue with Checkpoint restore( Beam pipeline) Date: Thu, 30 Nov 2017 10:48:03 +0100 In-Reply-To: <5A1F3A25.3070003@aeris.net> Cc: user@flink.apache.org To: Jins George References: <5A1F3A25.3070003@aeris.net> X-Mailer: Apple Mail (2.3445.4.7) archived-at: Thu, 30 Nov 2017 09:48:10 -0000 --Apple-Mail=_5901C47F-0570-44FE-873A-010BBC902026 Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset=us-ascii Hi, I think you might be running into a problem that is hard to solve with = Flink 1.2 and Beam. As you noticed, it's a problem that Beam doesn't = assign UIDs to operators, which is a problem. Flink 1.3 and even more = Flink 1.4 are a bit more lenient in accepting changes to the graph, so = you might have better luck when trying it with that. Did you try using a = newer Beam/Flink version? Flink 1.4 should be out next week and shortly = after that I'll also update the Beam dependency. Best, Aljoscha=20 > On 29. Nov 2017, at 23:52, Jins George wrote: >=20 > Hi, >=20 > I am running a Beam Pipeline on Flink 1.2 and facing an issue in = restoring a job from checkpoint. If I modify my beam pipeline to add a = new operator and try to restore from the externalized checkpoint, I get = the error=20 >=20 > java.lang.IllegalStateException: Invalid Invalid number of operator = states. Found :56. Expected: 58 > at = org.apache.flink.util.Preconditions.checkState(Preconditions.java:195) > at = org.apache.flink.streaming.runtime.tasks.StreamTask.checkRestorePreconditi= ons(StreamTask.java:680) > at = org.apache.flink.streaming.runtime.tasks.StreamTask.initializeState(Stream= Task.java:650) > at = org.apache.flink.streaming.runtime.tasks.StreamTask.invoke(StreamTask.java= :257) > at org.apache.flink.runtime.taskmanager.Task.run(Task.java:655) > at java.lang.Thread.run(Thread.java:745) >=20 > =46rom the savepoint guide [1], new operator added should be = initialized without any state. Any idea why this error is reported. >=20 > Also note, I am not setting a ID to my operator ( because Flink runner = in Beam does set the operator name user provided in pipeline creation) >=20 > [1] = https://ci.apache.org/projects/flink/flink-docs-release-1.3/setup/savepoin= ts.html = >=20 >=20 > Thanks, > Jins George --Apple-Mail=_5901C47F-0570-44FE-873A-010BBC902026 Content-Transfer-Encoding: quoted-printable Content-Type: text/html; charset=us-ascii Hi,

I = think you might be running into a problem that is hard to solve with = Flink 1.2 and Beam. As you noticed, it's a problem that Beam doesn't = assign UIDs to operators, which is a problem. Flink 1.3 and even more = Flink 1.4 are a bit more lenient in accepting changes to the graph, so = you might have better luck when trying it with that. Did you try using a = newer Beam/Flink version? Flink 1.4 should be out next week and shortly = after that I'll also update the Beam dependency.

Best,
Aljoscha 

On 29. Nov 2017, at 23:52, Jins = George <jins.george@aeris.net> wrote:

=20 =20
Hi,

I am running a Beam Pipeline on Flink 1.2 and facing an issue in restoring a job from checkpoint. If I modify my beam pipeline to add a new operator and  try to restore from the externalized = checkpoint, I get the error

java.lang.IllegalStateException: Invalid Invalid = number of operator states. Found :56. Expected: 58
    at = org.apache.flink.util.Preconditions.checkState(Preconditions.java:195)=
    at = org.apache.flink.streaming.runtime.tasks.StreamTask.checkRestorePreconditi= ons(StreamTask.java:680)
    at = org.apache.flink.streaming.runtime.tasks.StreamTask.initializeState(Stream= Task.java:650)
    at = org.apache.flink.streaming.runtime.tasks.StreamTask.invoke(StreamTask.java= :257)
    at org.apache.flink.runtime.taskmanager.Task.run(Task.java:655)
    at = java.lang.Thread.run(Thread.java:745)

=46rom the savepoint guide [1], new operator added should be initialized without any state.  Any idea why this error is = reported.

Also note, I am not setting a ID to my operator ( because Flink runner in Beam does set the operator name user provided in pipeline creation)

[1]  https://ci.apache.org/projects/flink/flink-docs-release-1= .3/setup/savepoints.html


Thanks,
Jins George

= --Apple-Mail=_5901C47F-0570-44FE-873A-010BBC902026--