From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
---|---|
To: | Robert Treat <xzilla(at)users(dot)sourceforge(dot)net> |
Cc: | pg(at)fastcrypt(dot)com, Paul Thomas <paul(at)tmsl(dot)demon(dot)co(dot)uk>, Shridhar Daithankar <shridhar_daithankar(at)myrealbox(dot)com>, pgsql-general(at)postgreSQL(dot)org |
Subject: | Re: MySQL Gets Functions in Java - Enlightenment Please |
Date: | 2003-12-22 16:02:15 |
Message-ID: | 16360.1072108935@sss.pgh.pa.us |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-general |
Robert Treat <xzilla(at)users(dot)sourceforge(dot)net> writes:
> On Sat, 2003-12-20 at 13:04, Dave Cramer wrote:
>> Is it more desirable to have a single java vm and communicate via RPC,
>> or some other mechanism?
> If you fire up the jvm at connection creation time rather than at first
> pljava function call, wouldn't that make things much simpler and faster
> per process?
You could possibly support both approaches by using the "library preload"
mechanism that Joe Conway developed recently. That is, the DBA can
either:
* Preload pljava, whereupon a JVM is created by the postmaster, and the
connection to this JVM is passed down to each backend and shared by
them.
* Do nothing, whereupon each backend that actually makes any pljava
function calls would have to start a JVM upon its first call.
I could see either of these approaches being the more efficient,
depending on how much pljava work is done in a given installation.
So being able to support both seems attractive.
regards, tom lane
From | Date | Subject | |
---|---|---|---|
Next Message | Keith C. Perry | 2003-12-22 16:16:47 | Re: windows distribution |
Previous Message | Sai Hertz And Control Systems | 2003-12-22 15:58:23 | Unix 2038 And PostgreSQL |