Mailing-List: contact user-help@cassandra.apache.org; run by ezmlm
Precedence: bulk
Reply-To: user@cassandra.apache.org
Received-SPF: pass (nike.apache.org: local policy)
DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws;
  s=s1024; d=yahoo.com;
  h=X-YMail-OSG:Received:X-Rocket-MIMEInfo:X-Mailer:Message-ID:Date:From:Reply-To:Subject:To:MIME-Version:Content-Type;
  b=QvDTNW1CnwClUVwQ34ob6Mrm1i9eS2rtM/j9O9zYdI/8xNK+97OkjKQkikmmS4jDynyh0jzc+c+UQeZ58sE6ejg0UOfk3h5T81EnUobJUjjfy3F0h6Y4q141nFiC3Y+5BGMdnx8MEK7cqcSa+KHQ75ea+BlPApnb8k5u1V2Tlfc=;
Message-ID: <1372358454.85687.YahooMailNeo@web142804.mail.bf1.yahoo.com>
Date: Thu, 27 Jun 2013 11:40:54 -0700 (PDT)
From: Bao Le <lvqb@yahoo.com>
Reply-To: Bao Le <lvqb@yahoo.com>
Subject: High-read latency for non-existing rows with LCS and 1.2.5
To: user@cassandra.apache.org
MIME-Version: 1.0
Content-Type: multipart/alternative;
 boundary="-2129327256-264758783-1372358454=:85687"

---2129327256-264758783-1372358454=:85687
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: quoted-printable

Hi,=0A=0A=A0=A0=A0 We are using Leveled Compaction with Cassandra 1.2.5. Ou=
r sstable size is 100M. On each node,=0Awe have anywhere from 700+ to 800+ =
sstables (for all levels). The bloom_filter_fp_chance is set at 0.000744.=
=0A=0A=A0 For read requests that ask for existing row records, the latency =
is great, mostly under 20 milliseconds with key cache and row cache set. Fo=
r read requests that ask for non-existing row records (not because of delet=
es, but rather, have never been in the system to start with), the latency i=
s running into hundreds of milliseconds and sometimes seconds.=0A=0A=A0 Jus=
t wonder if anyone has come across this before and has some pointers on how=
 to reduce the latency in this case.=0A=0AThanks=0ABao
---2129327256-264758783-1372358454=:85687
Content-Type: text/html; charset=iso-8859-1
Content-Transfer-Encoding: quoted-printable

<html><body><div style=3D"color:#000; background-color:#fff; font-family:ti=
mes new roman, new york, times, serif;font-size:12pt">Hi,<br><br>&nbsp;&nbs=
p;&nbsp; We are using Leveled Compaction with Cassandra 1.2.5. Our sstable =
size is 100M. On each node,<br>we have anywhere from 700+ to 800+ sstables =
(for all levels). The bloom_filter_fp_chance is set at 0.000744.<br><br>&nb=
sp; For read requests that ask for existing row records, the latency is gre=
at, mostly under 20 milliseconds with key cache and row cache set. For read=
 requests that ask for non-existing row records (not because of deletes, bu=
t rather, have never been in the system to start with), the latency is runn=
ing into hundreds of milliseconds and sometimes seconds.<br><br>&nbsp; Just=
 wonder if anyone has come across this before and has some pointers on how =
to reduce the latency in this case.<br><br>Thanks<br>Bao<br><br><br><div><b=
r></div></div></body></html>
---2129327256-264758783-1372358454=:85687--