Mailing-List: contact user-help@cassandra.apache.org; run by ezmlm
Precedence: bulk
Reply-To: user@cassandra.apache.org
Received-SPF: pass (athena.apache.org: domain of vram.kouramajian@gmail.com
 designates 209.85.212.44 as permitted sender)
DomainKey-Signature: a=rsa-sha1; c=nofws;
        d=gmail.com; s=gamma;
        h=mime-version:in-reply-to:references:date:message-id:subject:from:to
         :content-type;
        b=O8t81UTiQjfq+UhTruXE/qA3Iun0FiYF8cdXTI2GflIYhi+gYtALZs8bJAil3iFvsT
         /zEdxO4DYr5pqkiD3LAUG/ZZhLlsoIwLjC2Ft6oX4e6rav0Olxajr5ZEi091YZR4yNCB
         u213eNUnsiluPWu+EN2A0cm2KEKWoediRjiOI=
MIME-Version: 1.0
In-Reply-To: <8C1188E3-0B05-49EB-B512-BDF6344A6E3E@pagerduty.com>
References: <8C1188E3-0B05-49EB-B512-BDF6344A6E3E@pagerduty.com>
Date: Sat, 26 Jun 2010 15:24:00 -0700
Message-ID: <AANLkTinAb38ixb44Hd3b48zxhtRS_Iya_YwNHiEvOYAp@mail.gmail.com>
Subject: Re: Distributed work-queues?
From: Vram Kouramajian <vram.kouramajian@gmail.com>
To: user@cassandra.apache.org
Content-Type: multipart/alternative; boundary=0016e6469bc825a5540489f658e6

--0016e6469bc825a5540489f658e6
Content-Type: text/plain; charset=ISO-8859-1

We have implemented a distributed queue (similar to AWS SQS)  and a job
queue in Cassandra.

Vram


On Sat, Jun 26, 2010 at 1:56 PM, Andrew Miklas <andrew@pagerduty.com> wrote:

> Hi all,
>
> Has anyone written a work-queue implementation using Cassandra?
>
> There's a section in the UseCase wiki page for "A distributed Priority Job
> Queue" which looks perfect, but unfortunately it hasn't been filled in yet.
> http://wiki.apache.org/cassandra/UseCases#A_distributed_Priority_Job_Queue
>
> I've been thinking about how best to do this, but every solution I've
> thought of seems to have some serious drawback.  The "range ghost" problem
> in particular creates some issues.  I'm assuming each job has a row within
> some column family, where the row's key is the time at which the job should
> be run.  To find the next job, you'd do a range query with a start a few
> hours in the past, and an end at the current time.  Once a job is completed,
> you delete the row.
>
> The problem here is that you have to scan through deleted-but-not-yet-GCed
> rows each time you run the query.  Is there a better way?
>
> Preventing more than one worker from starting the same job seems like it
> would be a problem too.  You'd either need an external locking manager, or
> have to use some other protocol where workers write their ID into the row
> and then immediately read it back to confirm that they are the owner of the
> job.
>
> Any ideas here?  Has anyone come up with a nice implementation?  Is
> Cassandra not well suited for queue-like tasks?
>
>
>
> Thanks,
>
>
> Andrew
>

--0016e6469bc825a5540489f658e6
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

<br><br>We have implemented a distributed queue  (similar to AWS SQS)=A0 an=
d a job queue in Cassandra.<br><br>Vram<br><br><br><div class=3D"gmail_quot=
e">On Sat, Jun 26, 2010 at 1:56 PM, Andrew Miklas <span dir=3D"ltr">&lt;<a =
href=3D"mailto:andrew@pagerduty.com">andrew@pagerduty.com</a>&gt;</span> wr=
ote:<br>
<blockquote class=3D"gmail_quote" style=3D"margin: 0pt 0pt 0pt 0.8ex; borde=
r-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;">Hi all,<br>
<br>
Has anyone written a work-queue implementation using Cassandra?<br>
<br>
There&#39;s a section in the UseCase wiki page for &quot;A distributed Prio=
rity Job Queue&quot; which looks perfect, but unfortunately it hasn&#39;t b=
een filled in yet.<br>
<a href=3D"http://wiki.apache.org/cassandra/UseCases#A_distributed_Priority=
_Job_Queue" target=3D"_blank">http://wiki.apache.org/cassandra/UseCases#A_d=
istributed_Priority_Job_Queue</a><br>
<br>
I&#39;ve been thinking about how best to do this, but every solution I&#39;=
ve thought of seems to have some serious drawback. =A0The &quot;range ghost=
&quot; problem in particular creates some issues. =A0I&#39;m assuming each =
job has a row within some column family, where the row&#39;s key is the tim=
e at which the job should be run. =A0To find the next job, you&#39;d do a r=
ange query with a start a few hours in the past, and an end at the current =
time. =A0Once a job is completed, you delete the row.<br>

<br>
The problem here is that you have to scan through deleted-but-not-yet-GCed =
rows each time you run the query. =A0Is there a better way?<br>
<br>
Preventing more than one worker from starting the same job seems like it wo=
uld be a problem too. =A0You&#39;d either need an external locking manager,=
 or have to use some other protocol where workers write their ID into the r=
ow and then immediately read it back to confirm that they are the owner of =
the job.<br>

<br>
Any ideas here? =A0Has anyone come up with a nice implementation? =A0Is Cas=
sandra not well suited for queue-like tasks?<br>
<br>
<br>
<br>
Thanks,<br><font color=3D"#888888">
<br>
<br>
Andrew<br>
</font></blockquote></div><br>

--0016e6469bc825a5540489f658e6--