<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=us-ascii">
<meta name="GENERATOR" content="MSHTML 9.00.8112.16496">
</head>
<body>
<div dir="ltr" align="left"><span class="818123105-26082013"><font color="#0000ff" size="2" face="Arial">Hi Dan,</font></span></div>
<div dir="ltr" align="left"><span class="818123105-26082013"><font color="#0000ff" size="2" face="Arial"> </font></span></div>
<div dir="ltr" align="left"><span class="818123105-26082013"><font color="#0000ff" size="2" face="Arial"> Thanks for the information.</font></span></div>
<div dir="ltr" align="left"><span class="818123105-26082013"><font color="#0000ff" size="2" face="Arial"> Am using implicit transactions only..</font></span></div>
<div dir="ltr" align="left"><span class="818123105-26082013"><font color="#0000ff" size="2" face="Arial"> And i am on Infinispan - 5.3.0.</font></span></div>
<div dir="ltr" align="left"><span class="818123105-26082013"><font color="#0000ff" size="2" face="Arial"></font></span> </div>
<div dir="ltr" align="left"><span class="818123105-26082013"><font color="#0000ff" size="2" face="Arial">Thanks,</font></span></div>
<div dir="ltr" align="left"><span class="818123105-26082013"><font color="#0000ff" size="2" face="Arial">Faseela</font></span></div>
<br>
<div dir="ltr" lang="en-us" class="OutlookMessageHeader" align="left">
<hr tabindex="-1">
<font size="2" face="Tahoma"><b>From:</b> infinispan-dev-bounces@lists.jboss.org [mailto:infinispan-dev-bounces@lists.jboss.org]
<b>On Behalf Of </b>Dan Berindei<br>
<b>Sent:</b> Friday, August 23, 2013 6:49 PM<br>
<b>To:</b> infinispan -Dev List<br>
<b>Subject:</b> Re: [infinispan-dev] Eventual Consistency in Infinispan<br>
</font><br>
</div>
<div></div>
<div dir="ltr">
<div>
<div>
<div>
<div>Are you using explicit transactions or only implicit transactions?<br>
</div>
<br>
</div>
<div>As of 5.2, non-transactional caches use a scheme where the modification is sent to the primary owner, which then sends the modification to the backup owner(s). The primary owner waits for responses from all the backup owners, and the originator waits for
a response from the primary owner.<br>
<br>
</div>
<div>In optimistic caches, one-phase transactions are much simpler: the originator just sends the modification to all the owners, and they all apply it. Execution order is preserved for transactions originating from the same node, but transactions originating
from different nodes can be applied by the owners in any order.<br>
<br>
</div>
So implicit transactions with use1PcForAutoCommitTransactions="true" can certainly be faster than non-transactional operations, but I wouldn't recommend using them because they aren't atomic. (To be fair, there are also some problems with atomicity of non-tx
operations during state transfer that we're just fixing now - e.g. <a href="https://issues.jboss.org/browse/ISPN-3357">
https://issues.jboss.org/browse/ISPN-3357</a>)<br>
<br>
</div>
Cheers<br>
</div>
Dan<br>
<br>
</div>
<div class="gmail_extra"><br>
<br>
<div class="gmail_quote">On Mon, Aug 19, 2013 at 4:47 PM, Faseela K <span dir="ltr">
<<a href="mailto:faseela.k@ericsson.com" target="_blank">faseela.k@ericsson.com</a>></span> wrote:<br>
<blockquote style="BORDER-LEFT: #ccc 1px solid; MARGIN: 0px 0px 0px 0.8ex; PADDING-LEFT: 1ex" class="gmail_quote">
Hi Tristan,<br>
<br>
I was just comparing the performance of synchronous transactional and synchronous non-transactional writes in replication mode.<br>
And I see, transactional writes were taking lesser time, than non-transactional.<br>
Is it expected behaviour? My transactional config has the following properties :<br>
<br>
syncRollbackPhase="true"<br>
syncCommitPhase="true"<br>
cacheStopTimeout="30000"<br>
use1PcForAutoCommitTransactions="true"<br>
autoCommit="true"<br>
lockingMode="OPTIMISTIC"<br>
useSynchronization="true"<br>
transactionMode="TRANSACTIONAL"<br>
<br>
Or Am I using some wrong configuration?<br>
<br>
Thanks,<br>
Faseela<br>
<div class="HOEnZb">
<div class="h5"><br>
-----Original Message-----<br>
From: Tristan Tarrant [mailto:<a href="mailto:ttarrant@redhat.com">ttarrant@redhat.com</a>]<br>
Sent: Monday, August 19, 2013 6:28 PM<br>
To: infinispan -Dev List<br>
Cc: Faseela K<br>
Subject: Re: [infinispan-dev] Eventual Consistency in Infinispan<br>
<br>
Infinispan does not implement eventual consistency yet.<br>
<br>
Tristan<br>
<br>
On 08/19/2013 02:39 PM, Faseela K wrote:<br>
> Hi,<br>
> I am using Infinispan - 5.3.0 in clustering mode.<br>
> I have a transactional configuration, in replication-synchronous mode.<br>
> I want to know, whether eventual consistency is supported for<br>
> synchronous replication in 5.3.<br>
> Could someone please brief how eventual consistency works in<br>
> infinispan context?<br>
> Thanks,<br>
> Faseela<br>
><br>
><br>
> _______________________________________________<br>
> infinispan-dev mailing list<br>
> <a href="mailto:infinispan-dev@lists.jboss.org">infinispan-dev@lists.jboss.org</a><br>
> <a href="https://lists.jboss.org/mailman/listinfo/infinispan-dev" target="_blank">
https://lists.jboss.org/mailman/listinfo/infinispan-dev</a><br>
<br>
<br>
_______________________________________________<br>
infinispan-dev mailing list<br>
<a href="mailto:infinispan-dev@lists.jboss.org">infinispan-dev@lists.jboss.org</a><br>
<a href="https://lists.jboss.org/mailman/listinfo/infinispan-dev" target="_blank">https://lists.jboss.org/mailman/listinfo/infinispan-dev</a><br>
</div>
</div>
</blockquote>
</div>
<br>
</div>
</body>
</html>