Mailing-List: contact dev-help@jackrabbit.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@jackrabbit.apache.org
Received-SPF: pass (athena.apache.org: local policy)
From: =?iso-8859-1?Q?K=D6LL_Claus?= <C.KOELL@TIROL.GV.AT>
To: "dev@jackrabbit.apache.org" <dev@jackrabbit.apache.org>
Date: Mon, 25 Jun 2012 08:12:15 +0200
Subject: XA help needed ...
Thread-Topic: XA help needed ...
Thread-Index: Ac1SmCOAJ44xTuvzTquzmNcgbRYK9w==
Message-ID: <89934884426A01458CBE55947B042D7223B6234A60@EXCHMCA.tirol.local>
Accept-Language: de-DE, de-AT
Content-Language: de-DE
acceptlanguage: de-DE, de-AT
Content-Type: text/plain; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0

Hi ...

I found a "problem" within a xa environment and i don't know how to solve i=
t or=20
if it is possible to solve.

We have two J2EE Applications runnig on different AppServers and both have =
a Jackrabbit=20
Cluster Member instance running via JCA (as XAResource).
The Jackrabbit Cluster is configured to use a DB Journal.

So the problem is now if App1 adds a Node, then calls App2 and this also ad=
d a Node inside a global Transaction.

While preparing the global Transaction, JCR on App1 creates a global Lock o=
n DB (Global Revision Table). App2 tries also to prepare the transaction
but it can't add a DB Lock because App1 holds it now. App1 would free the L=
ock on commit ... so we have a global Deadlock.

Is it possible to create the Lock on the Global Revison Table only in one p=
hase (commit phase) and not spanned over prepare and commit ?

thanks for any help
claus