Daniel,
On 8/11/25 9:38 PM, Daniel Schwartz wrote:
Chris,
Again, see my replies marked DSG.
Dan
-----Original Message-----
From: Christopher Schultz <ch...@christopherschultz.net>
Sent: Monday, August 11, 2025 3:21 PM
To: users@tomcat.apache.org
Subject: Re: [EXTERNAL EMAIL] RE: How to access a REST service
Daniel,
On 8/5/25 3:05 PM, Daniel Schwartz wrote:
I just replied to Rob Sargent regarding this. Did you see that reply?
> Maybe I’m not using Glassfish correctly, but if I set the maximum> pool
size to 1, the system almost immediately outputs an error
message saying that it is unable to allocate any more connections and
crashes.
I'm interested in what you mean by "crashes". To me, that usually means that
the process stops at the very least. Do you mean that the request fails, or something
much worse?
DGS: By "crashes" I mean that the program terminates, usually just after
printing out a stack trace.
O_O
Like... the JVM completely exits and you have to re-start Tomcat? That
sounds ... suspicious. I don't think I've ever taken down Tomcat in the
20+ years I've been using it. Application in an unusable state? Sure.
But exiting the JVM usually requires a JVM bug or tcnative crash.
That’s why I have set a large maximum pool size.
I think you are fixing the symptom and not the problem. Even with a pool size
of 1, your server should not crash. The worst thing that should happen is a
request fails because a connection could not be immediately obtained from the
pool.
Most pools will wait some amount of time for a connection, and then time-out
and the request fails. I suppose it's possible to configure a pool to require a
connection to be immediately available, but that would be an unusual
configuration to use IMHO.
DGS: According to the people at Onmifish, if you set the maximum pool size to
1, you can have only one connection at a time.
This is exactly the intent of my recommendation.
So, if there is a request for a second connection, you get that message saying
that no more connections can be created, and the system crashes (stops
running). The behavior that you describe makes sense (sounds like how things
should work), but that evidently isn't true about Glassfish.
Maybe Glassfish's default configuration is to crap itself if you run out
of connections (which is *insane* IMHO). There must be a way to
configure that connection pool to "wait up to 5 seconds for a
connection, then throw an exception".
-chris
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org
For additional commands, e-mail: users-h...@tomcat.apache.org
