Mailing-List: contact user-help@zookeeper.apache.org; run by ezmlm
Precedence: bulk
Reply-To: user@zookeeper.apache.org
Message-ID: <COL126-W300517B37B4C6250E40C67B99F0@phx.gbl>
Content-Type: multipart/alternative;
	boundary="_bfb374e0-1916-4440-a3f4-b59c114a3a90_"
From: "snair123 ." <nairsom@outlook.com>
To: "user@zookeeper.apache.org" <user@zookeeper.apache.org>
Subject: RE: Tracking down possible network partition
Date: Fri, 10 Jul 2015 18:21:20 +0000
Importance: Normal
In-Reply-To: 
 <CAMUR=WjvX2Y7MpWm9CwbZsK70uN22ggk6TshvuvhmmemNW2q2w@mail.gmail.com>
References: 
 <D1B1CEA9.5F9C0%mark.round@bskyb.com>,<CAMUR=WjvX2Y7MpWm9CwbZsK70uN22ggk6TshvuvhmmemNW2q2w@mail.gmail.com>
MIME-Version: 1.0

--_bfb374e0-1916-4440-a3f4-b59c114a3a90_
Content-Type: text/plain; charset="Windows-1252"
Content-Transfer-Encoding: quoted-printable

2.) It appears that the leader closes connections to the affected followers
 after a =93transaction timeout=94 occurs. Where would I find out what this
 timeout is ? Is this the same thing as a session timout (e.g. The default
 of 20 * tickTime) ?

=20
 https://github.com/apache/zookeeper/blob/trunk/src/java/main/org/apache/zo=
okeeper/server/quorum/LearnerHandler.java#L496

a. So the Leader closes connections to Followers and Observers after syncLi=
mit*ticktime milliseconds ?=20
b. So what purpose does the SyncLimit serve in followers and in observers ?
c. If i needed the Observer to stay connected to ZKEnsemble for a longer ti=
me - in case of network partitiions - do i increase the syncLimit at the le=
ader or at the Observer ?


> Date: Fri=2C 26 Jun 2015 18:10:45 -0700
> Subject: Re: Tracking down possible network partition
> From: rgs@itevenworks.net
> To: user@zookeeper.apache.org
>=20
> On 25 June 2015 at 07:28=2C Round=2C Mark <Mark.Round@sky.uk> wrote:
>=20
> > I have a 5-node Zookeeper 3.4.6 cluster across 3 data centres (2
> > zookeepers in each =93main=94 DC=2C and a 5th in a 3rd DC for quorum). =
I see that
> > the two nodes in one DC have regular =93issues=94 where they get kicked=
 out of
> > the cluster and the ZooKeeperServer process stops for a few minutes unt=
il
> > the node rejoins. I=92d like to know a couple of things=2C if someone c=
ould
> > please point me in the direction of the relevant docs I=92d greatly
> > appreciate it.
> >
> > 1.) Is it expected behaviour that when a node is kicked from the cluste=
r=2C
> > it will not be allowed to re-join for a period ? From the logs below I =
can
> > see that re-establishing a valid cluster took around 15 minutes.
> >
>=20
> I don't think so.
>=20
> 2.) It appears that the leader closes connections to the affected followe=
rs
> > after a =93transaction timeout=94 occurs. Where would I find out what t=
his
> > timeout is ? Is this the same thing as a session timout (e.g. The defau=
lt
> > of 20 * tickTime) ?
> >
>=20
> https://github.com/apache/zookeeper/blob/trunk/src/java/main/org/apache/z=
ookeeper/server/quorum/LearnerHandler.java#L496
>=20
>=20
> > 3.) Where can I find the definition of the different fields in the
> > election log messages (I.e. What are =93n.round=94=2C =93n.zxid=94=2C =
=93n.state=94 and so
> > on) ?
>=20
>=20
> Not sure if there's a better source than the source:
> https://github.com/apache/zookeeper/blob/trunk/src/java/main/org/apache/z=
ookeeper/server/quorum/FastLeaderElection.java#L687
>=20
>=20
>=20
> -rgs
 		 	   		  =

--_bfb374e0-1916-4440-a3f4-b59c114a3a90_--