From: thomas at tada.se (Thomas Hallgren) Date: Mon, 13 Dec 2010 15:26:42 +0100 Subject: [Pljava-dev] pljava @ itpug (italy) 2010 In-Reply-To: <201012131437.12442.fluca1978@infinito.it> References: <201012130834.13591.fluca1978@infinito.it> <201012131437.12442.fluca1978@infinito.it> Message-ID: <4D062D22.8090409@tada.se> Hi Luca, Wouldn't it be better to manage distributed transactions in the middle tier on top of it all? After all, you are already *in* a transaction when you execute your PL/Java code and that's a transaction that you cannot connect other resources to. When writing PL/Java, I did not intend for distributed transactions to be orchestrated from within it for sure. That's not a recommended use-case. In fact, I would strongly advice against it. Regards, Thomas Hallgren On 2010-12-13 14:37, Luca Ferrari wrote: > On Monday, December 13, 2010 01:57:48 pm Johann 'Myrkraverk' Oskarsson's cat > walking on the keyboard wrote: > >> Could you clarify a little, I'm not following. PL/Java runs in the >> backend already. >> > > Let say pl7java is executing in a transaction scope on the postgresql server, > and connects/starts a distributed transaction on another application server. > Such transaction is totally managed by the pl/java process, but the backend > does not see it at all, right? Is there a way to make the distributed > transaction started from pl/java available on the backed? So if the former > aborts the second does too. My quick answer is no, but I ask to be sure. > > Thanks, > Luca > _______________________________________________ > Pljava-dev mailing list > Pljava-dev at pgfoundry.org > http://pgfoundry.org/mailman/listinfo/pljava-dev