My deployments are Websphere deployments and I use the connection pooling of
Websphere, otherwise I would be an enemy of myself in case of
troubleshooting with the IBM Websphere support.
I am really neutral on this and I do not use DBCP.
Andrea Aime writes:
> Justin Deoliveira ha scritto
Justin Deoliveira ha scritto:
> I would be fine with straight switching. It seems we don't lose any
> functionality and if anything we gain some performance. Any down side?
>
> I guess it might be nice to give the user the choice, allowing them to
> swap but not sure how much it gains us. And th
I would be fine with straight switching. It seems we don't lose any
functionality and if anything we gain some performance. Any down side?
I guess it might be nice to give the user the choice, allowing them to
swap but not sure how much it gains us. And they can fall back on JNDI
in such cases.
Andrea Aime ha scritto:
> Hi,
> I've been experimenting a little with connection pools performance
> and it seems the C3P0 connection pool
> (http://sourceforge.net/projects/c3p0/)
> is designed for slightly better scalability than the DBCP one we're
> using today.
>
> In particular in a WMS requ
Andrea Aime wrote:
> Ben Caradoc-Davies ha scritto:
>> This is a documentation issue. Decent examples, such as Andrea's JNDI
>> tutorial, make things much easier. I spent more time getting JNDI to
>> work with embedded Jetty.
>
> Hmmm... I guess when we think about "user" we think different thin
Ben Caradoc-Davies ha scritto:
> Christian Müller wrote:
>> I think that all deployments in a J2EE container should use the
>> connection pooling of the chosen container.
>
> Christian, as reminded me, GeoTools datastores may be used without a
> J2EE container. I think this is what Andrea is pr
Ben Caradoc-Davies ha scritto:
> Andrea Aime wrote:
>> Simple: setting up a JNDI connection pool is _hard_ for most of the
>> people (go ask Ben how much time he wasted setting it up, and
>> I can tell you I cursed Tomcat for a good two hours before I got my
>> JNDI pool working).
>
> This is a do
Christian Müller wrote:
> I think that all deployments in a J2EE container should use the connection
> pooling of the chosen container.
Christian, as reminded me, GeoTools datastores may be used without a
J2EE container. I think this is what Andrea is proposing: a change to
the connection pool
Andrea Aime wrote:
> Simple: setting up a JNDI connection pool is _hard_ for most of the
> people (go ask Ben how much time he wasted setting it up, and
> I can tell you I cursed Tomcat for a good two hours before I got my
> JNDI pool working).
This is a documentation issue. Decent examples, such
Christian Müller ha scritto:
> I think that all deployments in a J2EE container should use the
> connection pooling of the chosen container.
> I do not see the importance of a change but perhaps I have overseen
> something. ?
Simple: setting up a JNDI connection pool is _hard_ for most of the
pe
My 2 cents,
DBCP is not usually recommended in the JEE environments where I have
worked (mainly JBOSS) while c3p0 is usually highly recommended.
Simone.
---
Ing. Simone Giannecchini
GeoSolutions S.A.S.
Owner - Software Engineer
Via Carignoni 51
5
I think that all deployments in a J2EE container should use the connection
pooling of the chosen container.
I do not see the importance of a change but perhaps I have overseen
something. ?
Andrea Aime writes:
> Hi,
> I've been experimenting a little with connection pools performance
> and i
Hi,
I've been experimenting a little with connection pools performance
and it seems the C3P0 connection pool
(http://sourceforge.net/projects/c3p0/)
is designed for slightly better scalability than the DBCP one we're
using today.
In particular in a WMS request benchmark, and when reaching the h
13 matches
Mail list logo