Return-Path: X-Original-To: apmail-cassandra-user-archive@www.apache.org Delivered-To: apmail-cassandra-user-archive@www.apache.org Received: from mail.apache.org (hermes.apache.org [140.211.11.3]) by minotaur.apache.org (Postfix) with SMTP id 94E7F10635 for ; Sat, 20 Jul 2013 16:03:17 +0000 (UTC) Received: (qmail 61037 invoked by uid 500); 20 Jul 2013 16:03:15 -0000 Delivered-To: apmail-cassandra-user-archive@cassandra.apache.org Received: (qmail 60742 invoked by uid 500); 20 Jul 2013 16:03:11 -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 60732 invoked by uid 99); 20 Jul 2013 16:03:10 -0000 Received: from nike.apache.org (HELO nike.apache.org) (192.87.106.230) by apache.org (qpsmtpd/0.29) with ESMTP; Sat, 20 Jul 2013 16:03:10 +0000 X-ASF-Spam-Status: No, hits=1.5 required=5.0 tests=HTML_MESSAGE,RCVD_IN_DNSWL_LOW,SPF_PASS X-Spam-Check-By: apache.org Received-SPF: pass (nike.apache.org: local policy) Received: from [128.210.5.246] (HELO mailhub246.itcs.purdue.edu) (128.210.5.246) by apache.org (qpsmtpd/0.29) with ESMTP; Sat, 20 Jul 2013 16:03:04 +0000 Received: from mail-vc0-f174.google.com (mail-vc0-f174.google.com [209.85.220.174]) (authenticated bits=0) by mailhub246.itcs.purdue.edu (8.14.4/8.14.4/mta-auth.smtp.purdue.edu) with ESMTP id r6KG2fRa013238 (version=TLSv1/SSLv3 cipher=RC4-SHA bits=128 verify=NOT) for ; Sat, 20 Jul 2013 12:02:41 -0400 Received: by mail-vc0-f174.google.com with SMTP id ht11so147337vcb.19 for ; Sat, 20 Jul 2013 09:02:40 -0700 (PDT) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20120113; h=mime-version:in-reply-to:references:from:date:message-id:subject:to :content-type; bh=TQ12CA60+24suba1JKC3alQ2sVYLHRVU23oevfbaIFw=; b=ACfQy7KClqHgeTUE3pAFcVI8sEJUGjgteTp5EQKcQk9uAZQLcGthCU30MeJFQPF7O0 a8mkD2NxCuM8ppQqj6odvCQjyXC9U+D0fHICmxZ45dMQH2tUI8EyIfsXNojJHGTmGkny lwdIFlHoZ9hj2LW8YNBzpXIBn35/c0ioX+w2KxR7iqt/n9/jesUTdIXAnQFLeJVHSLHO tHKeg2Mm5bqwa/r3DO4OBfirwK1CsWN19IbUkzgjYx0yeQg9aBLu9AuPcOUJOjXEkrV6 Mq3EIeR71Dt2e7uxqDXnUqVWD5nr/CB1xWPclPNDQjKLbe43sg6Of431HutyX0WXfWDs 3S9g== X-Received: by 10.221.4.4 with SMTP id oa4mr7208037vcb.70.1374336160946; Sat, 20 Jul 2013 09:02:40 -0700 (PDT) MIME-Version: 1.0 Received: by 10.58.229.164 with HTTP; Sat, 20 Jul 2013 09:02:00 -0700 (PDT) In-Reply-To: References: <57C7C3CBDCB04F45A57AEC4CB21C0CCD1DBDEEB0@mbx024-e1-nj-6.exch024.domain.local> From: Mohammad Hajjat Date: Sat, 20 Jul 2013 12:02:00 -0400 Message-ID: Subject: Re: CPU Bound Writes To: user@cassandra.apache.org Content-Type: multipart/alternative; boundary=089e01161992ace72a04e1f39229 X-PMX-Version: 6.0.2.2308539 X-PerlMx-Virus-Scanned: Yes X-Virus-Checked: Checked by ClamAV on apache.org --089e01161992ace72a04e1f39229 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: quoted-printable Patricia, Thanks for the info. So are you saying that the *whole* data is being written on disk in the commit log, not just some sort of a summary/digest? I'm writing 10MB objects and I'm noticing high latency (250 milliseconds even with ANY consistency), so I guess that explains my high delays? Thanks, Mohammad On Fri, Jul 19, 2013 at 2:17 PM, Patricia Gorla w= rote: > Kanwar, > > This is because writes are appends to the commit log, which is stored on > disk, not memory. The commit log is then flushed to the memtable (in > memory), before being written to an sstable on disk. > > So, most of the actions in sending out a write are writing to disk. > > Also see: http://www.datastax.com/docs/1.2/dml/about_writes > > Patricia > > > On Fri, Jul 19, 2013 at 1:05 PM, Kanwar Sangha wrote= : > >> =E2=80=9CInsert-heavy workloads will actually be CPU-bound in Cassandra= before >> being memory-bound=E2=80=9D**** >> >> ** ** >> >> Can someone explain why the internals of why writes are CPU bound ?**** >> >> ** ** >> >> ** ** >> > > --=20 *Mohammad Hajjat* *Ph.D. Student* *Electrical and Computer Engineering* *Purdue University* --089e01161992ace72a04e1f39229 Content-Type: text/html; charset=UTF-8 Content-Transfer-Encoding: quoted-printable
Patricia,=C2=A0

Thanks for the info. So are you saying that the *whole* data is being writt= en on disk in the commit log, not just some sort of a summary/digest?
=
I'= ;m writing 10MB objects and I'm noticing high latency (250 milliseconds= even with ANY consistency), so I guess that explains my high delays?

<= /div>
= Thanks,
Mohammad


On Fri, Jul 19, 2013 at 2:17 PM, Patricia Gorla <gorla.patricia@gm= ail.com> wrote:
Kanwar,

= This is because writes are appends to the commit log, which is stored on di= sk, not memory. The commit log is then flushed to the memtable (in memory),= before being written to an sstable on disk.

So, most of the actions in sending out a write are writ= ing to disk.


Patricia


On Fri, Jul 19, 2013 at 1:05 PM, Kanwar = Sangha <kanwar@mavenir.com> wrote:

=E2=80=9CInsert-heavy workloads= will actually be CPU-bound in Cassandra before being memory-bound=E2=80=9D=

=C2=A0

Can someone explain why the int= ernals of why writes are CPU bound ?

=C2=A0

=C2=A0





--
Mohammad Hajjat=
Ph.D. Student
Electrical and Computer Engineering
Purdue University
<= /font>
--089e01161992ace72a04e1f39229--