Subject | Re: [Firebird-Java] Release status |
---|---|
Author | William L. Thomson Jr. |
Post date | 2002-12-10T03:19:08Z |
A little off topic
On Mon, 2002-12-09 at 15:47, Rick Fincher wrote:
>
> but we've been running it with Tomcat on Solaris in a
> small company and JayBird is very solid. Much more so than Tomcat.
I second that.
> Firebird never crashes. With Interclient the web apps would occasionally
> get non-responsive and I'd have to restart Tomcat. I see much less of that
> with JayBird, and the stand alone apps that I have running JayBird have run
> flawlessly with the same databases, so I'm certain Tomcat is the culprit in
> any instability.
I will have to agree to Tomcat being imperfect. I have had many problems
like some sort of memory leak or something. I have gone through all my
code a thousand times, and many version of JayBird, mostly from CVS. At
this point I am pretty sure it's Tomcat fault since restarting Tomcat
always makes the weirdest of problems go away. Boy do I see some weird
ones with Tomcat. In my case my biggest problem is gc. I do not think it
is happening, or at least not often enough.
> To test that I ran Tomcat without making any database calls in the web apps
> and had the same instability, so while not absolute proof I'm convinced that
> JayBird is very stable.
I have been using the beta release is client apps, and they have not had
any problems to date. I know that there are bugs, and etc with the beta
driver, but since they are not causing problems I have not yet
distributed a replacement to my clients. I will when there is a final
release, or if a problem arises.
On the server side that has not been the case, and rather than blame
JayBird I am pretty sure it's Tomcat's fault.
> I don't have high loads so I can't tell you how JayBird would behave under
> stress, but from what I've seen in a servlet container environment the JDBC
> driver accounts for very little overhead. The performance of the app server
> and the database itself are far more of a factor.
I have decent volume on our production server and very little if any
problems related to JayBird. I have used a recent CVS version to
restructure a database storing 20-30kb pictues in blob format. I pumped
over 8,000 images in and out of the db. All in all took about 20
minutes. Which I do not think is bad for select then insert for the
8,000 pictures.
So in my opinion JayBird does fairly well with high volumes, or at least
the level of volume I am dealing with. I can provide detail information
on the specific amount of volume if it's of interest to anyone. Just
email me.
> I'm a little surprised about your experience with the high speed of JayBird.
> Your particular apps must really give Interclient fits. Not to belittle
> JayBird but most people have not seen huge boots in performance with
> JayBird.
I noticed a slight degradation in performance when I first switched from
InterClient to JayBird. For certain things it's still a little slower,
but for others JayBird is at least the same. Nothing seemed to be
faster. However I was able to get a little better performance due to
coding suggestions like use execute instead of executeUpdate. Fairly big
difference on WAN's, not so much on LAN's. Still worth modifying the
code.
> Do you think this will be true with the higher loads on your production
> server? It may be that Interserver becomes a bottleneck under certain
> circumstances.
Questions I would like to know the answer to as well. In my case our
loads are greater now then when using InterClient. In the back of my
mind I have always wondered what performance would be like if I went
back.
I won't go back because I feel it's my duty to use JayBird. Since people
are taking allot of their time to make it all it can be. I wish the same
could be said for InterClient in the open source world.
So even if JayBird is not as good as InterClient in one way or another,
I am not sure if it will always be that way.
If anything in the end JayBird will be better or the only one for
Firebird in a year or two. Which in the end is what was the deciding
factor to switch.
--
Sincerely,
William L. Thomson Jr.
Support Group
Obsidian-Studios Inc.
439 Amber Way
Petaluma, Ca. 94952
Phone 707.766.9509
Fax 707.766.8989
http://www.obsidian-studios.com