From user-return-9479-apmail-cassandra-user-archive=cassandra.apache.org@cassandra.apache.org Mon Sep 27 04:32:12 2010 Return-Path: Delivered-To: apmail-cassandra-user-archive@www.apache.org Received: (qmail 12124 invoked from network); 27 Sep 2010 04:32:11 -0000 Received: from unknown (HELO mail.apache.org) (140.211.11.3) by 140.211.11.9 with SMTP; 27 Sep 2010 04:32:11 -0000 Received: (qmail 97823 invoked by uid 500); 27 Sep 2010 04:32:10 -0000 Delivered-To: apmail-cassandra-user-archive@cassandra.apache.org Received: (qmail 97604 invoked by uid 500); 27 Sep 2010 04:32:06 -0000 Mailing-List: contact user-help@cassandra.apache.org; run by ezmlm Precedence: bulk List-Help: List-Unsubscribe: List-Post: List-Id: Reply-To: user@cassandra.apache.org Delivered-To: mailing list user@cassandra.apache.org Received: (qmail 97596 invoked by uid 99); 27 Sep 2010 04:32:05 -0000 Received: from nike.apache.org (HELO nike.apache.org) (192.87.106.230) by apache.org (qpsmtpd/0.29) with ESMTP; Mon, 27 Sep 2010 04:32:05 +0000 X-ASF-Spam-Status: No, hits=2.2 required=10.0 tests=HTML_MESSAGE,MIME_QP_LONG_LINE,RCVD_IN_DNSWL_NONE,SPF_PASS X-Spam-Check-By: apache.org Received-SPF: pass (nike.apache.org: local policy) Received: from [208.113.200.5] (HELO homiemail-a59.g.dreamhost.com) (208.113.200.5) by apache.org (qpsmtpd/0.29) with ESMTP; Mon, 27 Sep 2010 04:31:59 +0000 Received: from homiemail-a59.g.dreamhost.com (localhost [127.0.0.1]) by homiemail-a59.g.dreamhost.com (Postfix) with ESMTP id E62CC564069 for ; Sun, 26 Sep 2010 21:31:35 -0700 (PDT) DomainKey-Signature: a=rsa-sha1; c=nofws; d=thelastpickle.com; h=to:from :subject:date:message-id:content-type:mime-version:in-reply-to; q=dns; s=thelastpickle.com; b=0l+1lgeo0keeY+mmkbgT+7YsPE5q4atKz QEK6+xRx70/g/hi1u6Ljf0emdKJkM+3nmZ5ndeZl411pjGZ4fD/G7aofDbwLGH87 THoAHgjiHjhqi9skNUe3DN19yq3o2RT9kvudDW1pAUIo1loEzA6/r03g8t1qtIbJ W1diezfMX0= DKIM-Signature: v=1; a=rsa-sha1; c=relaxed; d=thelastpickle.com; h=to :from:subject:date:message-id:content-type:mime-version: in-reply-to; s=thelastpickle.com; bh=RbOgpMBTqfAntg4TSrEg+/rEqag =; b=ZZAS2GMwNKwEafk0FuFAwtTqJRALljQcZ6a6wmE0V4pT6871c007HYEgSpG FpdP+RQ+phLX0G82Tvc6znPAAdF2lLVcMdkMVnqZBZSAweo8VPdBXDriadaCBPUJ rb8dqmVdNlQAuRcv69HMC7HOF6zNb4vLBvnmqKvgi8iPJJPI= Received: from localhost (webms.mac.com [17.148.16.116]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) (Authenticated sender: aaron@thelastpickle.com) by homiemail-a59.g.dreamhost.com (Postfix) with ESMTPSA id CFA87564061 for ; Sun, 26 Sep 2010 21:31:35 -0700 (PDT) To: user@cassandra.apache.org From: Aaron Morton Subject: Re: inter node protocol for 0.7 nightly Date: Mon, 27 Sep 2010 04:31:34 GMT X-Mailer: MobileMe Mail (1C3203) Message-id: <5916bc8e-c9a5-49a4-a085-d788943b28a1@me.com> Content-Type: multipart/alternative; boundary=Apple-Webmail-42--f14a17a6-ad1f-6464-a6ed-e63e8e06fad7 MIME-Version: 1.0 In-Reply-To: X-Virus-Checked: Checked by ClamAV on apache.org --Apple-Webmail-42--f14a17a6-ad1f-6464-a6ed-e63e8e06fad7 Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset=ISO-8859-1; format=flowed Yeah, I can only get away with it because the nodes have only 20GB each an= d there's a load of free space.=A0=0A=0AThanks=0AAaron=0A=0A=0AOn 27 Sep, = 2010,at 05:25 PM, Jonathan Ellis wrote:=0A=0Athat shou= ld work, but requiring RF=3DN makes it of limited usefulness in=0Apractice= :)=0A=0AOn Sun, Sep 26, 2010 at 9:10 PM, Aaron Morton wrote:=0A> It is indeed CASSANDRA-1465 where the NEWS.TXT was chang= ed to say the wire=0A> protocol had changed=0A> http://github.com/apache/c= assandra/commit/4023c3b6f9d4cd66d56024b07962968f2424815f#diff-1=0A> As far= as I can see the wire protocol change is to remove the string name of=0A>= the TP stage from the message. I can do a shutdown upgrade, but was think= ing=0A> of trying the following just to see if it works.=0A> 1) I have a 4= node with RF3, first increase the RF to 4 and repair so each=0A> node has= all the=0A> data=A0http://wiki.apache.org/cassandra/FAQ#change_replicatio= n=0A> 2) change my clients to use CL.ANY (i guess ONE would work as well)=0A= > 3) change the listen_address on the nodes to localhost=0A> What I think = I've done now is make a cluster of individual nodes, that=0A> cannot talk = to each other. When a client connects (it has a list of nodes)=0A> to one = node it should be able to read data and perform writes. HH's will be=0A> s= tored to send the data to the other nodes in the cluster.=0A> 4) Upgrade e= ach node in turn to the new 0.7 nightly, I would drain the node=0A> and de= lete the schema CF's as=0A> per=A0http://www.mail-archive.com/user@cassand= ra.apache.org/msg05726.html=0A> 5) return the listen_address back to an em= pty setting, so the nodes see each=0A> other again. Then run repair on eac= h node in turn to deliver the HH they=0A> collected while they could not c= ommunicate.=0A> 6)=A0Change the RF back to 3 and nodetool cleanup on each=0A= > Aside from "huh why?" any thoughts on if this would work? I have a sligh= t=0A> opportunity to play around and am interested to see if I can roll th= is out=0A> without a full shutdown.=0A> Thanks=0A> Aaron=0A>=0A> On 22 Sep= , 2010,at 02:57 PM, Jonathan Ellis wrote:=0A>=0A> Yes,= I think that's the one.=0A>=0A> I imagine svn blame on NEWS would tell yo= u for sure.=0A>=0A> On Tue, Sep 21, 2010 at 8:05 AM, Gary Dusbabek wrote:=0A>> 1465 maybe?=0A>>=0A>> On Mon, Sep 20, 2010 at 1= 6:00, Aaron Morton =0A>> wrote:=0A>>> Just took a= look upgrading from from 31/08 nightly to the 20/09 and=0A>>> noticed=0A>= >> the news.txt says...=0A>>> "The Cassandra inter-node protocol is incomp= atible with 0.6.x=A0releases=0A>>> (and=0A>>> with 0.7 beta1)"=0A>>> Could= someone point me to the ticket(s) for this change so I can see if I=0A>>>= can do a rolling upgrade.=0A>>> Thanks=0A>>> Aaron=0A>>>=0A>>=0A>=0A>=0A>= =0A> --=0A> Jonathan Ellis=0A> Project Chair, Apache Cassandra=0A> co-foun= der of Riptano, the source for professional Cassandra support=0A> http://r= iptano.com=0A>=0A=0A=0A=0A-- =0AJonathan Ellis=0AProject Chair, Apache Cas= sandra=0Aco-founder of Riptano, the source for professional Cassandra supp= ort=0Ahttp://riptano.com=0A --Apple-Webmail-42--f14a17a6-ad1f-6464-a6ed-e63e8e06fad7 Content-Type: multipart/related; type="text/html"; boundary=Apple-Webmail-86--f14a17a6-ad1f-6464-a6ed-e63e8e06fad7 --Apple-Webmail-86--f14a17a6-ad1f-6464-a6ed-e63e8e06fad7 Content-Transfer-Encoding: quoted-printable Content-Type: text/html; charset=ISO-8859-1;
Yeah, I can only get away with it because the nodes have only 20GB ea= ch and there's a load of free space. 

Thanks=
Aaron


On 27 Sep, 2010,at 05:25 PM, Jonathan E= llis <jbellis@gmail.com> wrote:

that should work, but requiring RF=3DN= makes it of limited usefulness in
=0Apractice :)
=0A
=0AOn Sun, = Sep 26, 2010 at 9:10 PM, Aaron Morton <aaron@thelastpickle.com> wrot= e:
=0A> It is indeed CASSANDRA-1465 where the NEWS.TXT was changed t= o say the wire
=0A> protocol had changed
=0A> http://github.com/apache/cassand= ra/commit/4023c3b6f9d4cd66d56024b07962968f2424815f#diff-1
=0A> A= s far as I can see the wire protocol change is to remove the string name o= f
=0A> the TP stage from the message. I can do a shutdown upgrade, b= ut was thinking
=0A> of trying the following just to see if it works=
=0A> 1) I have a 4 node with RF3, first increase the RF to 4 and r= epair so each
=0A> node has all the
=0A> data http://wiki.apache.or= g/cassandra/FAQ#change_replication
=0A> 2) change my clients to = use CL.ANY (i guess ONE would work as well)
=0A> 3) change the liste= n_address on the nodes to localhost
=0A> What I think I've done now = is make a cluster of individual nodes, that
=0A> cannot talk to each= other. When a client connects (it has a list of nodes)
=0A> to one = node it should be able to read data and perform writes. HH's will be
=0A= > stored to send the data to the other nodes in the cluster.
=0A>= 4) Upgrade each node in turn to the new 0.7 nightly, I would drain the no= de
=0A> and delete the schema CF's as
=0A> per http://www.mail-archive.com/user@cassandra.apache.org/msg05726.html
=0A> 5) return the listen_address back to an empty setting, so the= nodes see each
=0A> other again. Then run repair on each node in tu= rn to deliver the HH they
=0A> collected while they could not commun= icate.
=0A> 6) Change the RF back to 3 and nodetool cleanup on = each
=0A> Aside from "huh why?" any thoughts on if this would work? = I have a slight
=0A> opportunity to play around and am interested to= see if I can roll this out
=0A> without a full shutdown.
=0A>= Thanks
=0A> Aaron
=0A>
=0A> On 22 Sep, 2010,at 02:57 PM= , Jonathan Ellis <jbellis@gmail.com> wrote:
=0A>
=0A> Ye= s, I think that's the one.
=0A>
=0A> I imagine svn blame on NE= WS would tell you for sure.
=0A>
=0A> On Tue, Sep 21, 2010 at = 8:05 AM, Gary Dusbabek <gdusbabek@gmail.com> wrote:
=0A>> 1= 465 maybe?
=0A>>
=0A>> On Mon, Sep 20, 2010 at 16:00, Aa= ron Morton <aaron@thelastpickle.com>
=0A>> wrote:
=0A>= ;>> Just took a look upgrading from from 31/08 nightly to the 20/09 = and
=0A>>> noticed
=0A>>> the news.txt says...
= =0A>>> "The Cassandra inter-node protocol is incompatible with 0.= 6.x releases
=0A>>> (and
=0A>>> with 0.7 beta= 1)"
=0A>>> Could someone point me to the ticket(s) for this ch= ange so I can see if I
=0A>>> can do a rolling upgrade.
=0A= >>> Thanks
=0A>>> Aaron
=0A>>>
=0A>= >
=0A>
=0A>
=0A>
=0A> --
=0A> Jonathan El= lis
=0A> Project Chair, Apache Cassandra
=0A> co-founder of Ri= ptano, the source for professional Cassandra support
=0A> http://riptano.com
=0A>
=0A
=0A
=0A
=0A--
=0AJonathan Ellis
=0APr= oject Chair, Apache Cassandra
=0Aco-founder of Riptano, the source for = professional Cassandra support
=0Ahttp://riptano.com
=0A
--Apple-Webmail-86--f14a17a6-ad1f-6464-a6ed-e63e8e06fad7-- --Apple-Webmail-42--f14a17a6-ad1f-6464-a6ed-e63e8e06fad7--