Mailing-List: contact user-help@zookeeper.apache.org; run by ezmlm
Precedence: bulk
Reply-To: user@zookeeper.apache.org
Received-SPF: pass (athena.apache.org: domain of prvs=9876185c63=thawan@fb.com
 designates 67.231.145.42 as permitted sender)
From: Thawan Kooburat <thawan@fb.com>
To: "user@zookeeper.apache.org" <user@zookeeper.apache.org>
Subject: Re: leader election length
Thread-Topic: leader election length
Thread-Index: AQHOaHWWbMoNdg0y10CLZ8RgfmbgSJk0KNqA
Date: Thu, 13 Jun 2013 21:28:04 +0000
Message-ID: 
 <AA0B8A9CF5974846889A3DF776C1DF17569B1C09@PRN-MBX02-2.TheFacebook.com>
In-Reply-To: <1371155701370-7578752.post@n2.nabble.com>
Accept-Language: en-US
Content-Language: en-US
Content-Type: text/plain; charset="us-ascii"
Content-ID: <D296FCC7BB5FFF45A1323CE4C4D5F0A5@fb.com>
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0

If your snapshot size wasn't big.  You can increase the snapshot frequency
(Java prop: zookeeper.snapCount)

This may help a bit in reducing the size of txnlog, so it take less time
for the server to load the data on leader election.

However, taking more frequent snapshot cause the server to use more disk
bandwidth during normal operation. It may affect request latency if both
txnlog and snapshot shares the same disk.

--=20
Thawan Kooburat


On 6/13/13 1:35 PM, "ankitoshniwal" <ankitoshniwal@gmail.com> wrote:

>Hello,
>
>We are running into a similar issues. Is there recommended path to take in
>this scenario. We are running version 3.4.4 of zookeeper and the txlog on
>disk is close to 5 - 6 Gb. Currently the leader election sometimes takes
>up
>to 50 seconds.
>
>Any help is greatly appreciated.
>
>Thanks,
>Ankit.
>
>
>
>--
>View this message in context:
>http://zookeeper-user.578899.n2.nabble.com/leader-election-length-tp708686
>8p7578752.html
>Sent from the zookeeper-user mailing list archive at Nabble.com.