Dear Seagate.
I'm not an expert on GF, but a frustrated user too. I feel the version I'm
using is not mature enough and I would use it in production environments with
caution.
I do agree that your words may impact as not respectfull, although
I understand they are born due to the frustration and non-sensical errors
you may see.
Still, the community is not to blame, we are here to help. And I would say
to community to please understand GF does have some problems, may be minor
ones, but it is not perfect.
Now, on your question. It is very difficult to determine what may be causing
the problems without more info, stack traces and such. As you may be hosting
many clients, you know that what one of them does may impact others. So, it
is important to check if the problem has a cause that is not totally related
to GF, or if there is something running that may be affecting it.
I had found that the console is a thing you should not touch much. It
trashes GF when I do a simple edit on the log config, and the solution is a
asadming call to restore the log configuration (a workaround). So, I suggest
to avoid the console and try to perform all the edits using CLI.
Memory, Well, you know, memory leaks may be at any level. JDK implementation,
server or even the apps that are running in it. So, again, check the cause,
try to perform a memory dump and analyse it to see what is eating all that
memory. We may find it is a client's app, or the client's app is causing GF
to use something with a memory leak. Once found, the cause can be removed by
either asking the client to fix the app or sending the info to the GF guys
and filing a bug.
Lastly, are you using a single instance to do all this? Have you tried
clustering? Even in the same machine you can have several nodes running. And,
you can also use the monitoring options to learn when the instance is about
to crack, and even detect what can be causing the problem.
Hope this helps.
William.
--
[Message sent by forum member 'wmartinez']
View Post: http://forums.java.net/node/835908