dev@glassfish.java.net

Re: Waiting 600,000 milliseconds to start DAS?

From: Bill Shannon <bill.shannon_at_oracle.com>
Date: Wed, 16 Jun 2010 20:27:31 -0700

The start-domain timeout is large because a large domain with lots of
applications can take a long time to start. If startup is hanging
instead of succeeding or failing, we should do what we can to figure
out why and prevent it or detect it.


Byron Nevins wrote on 06/14/2010 07:36 PM:
> I've seen this come up a few times. Developers sitting around waiting
> 10 minutes for a server to start.
>
> Don't do that!
>
> 2 issues:
>
> 1) GlassFish will happily hang instead of dying gracefully.
> 2) When (1) happens it is frequently very early in the startup process.
> Before the logging service starts. So nothing will be logged. E.g.,
> if anything goes wrong in Felix initialization you will never know about
> it if you start the server with "asadmin start-domain". Yes - that's
> right. All messages "logged" before the Logging Service starts are
> thrown away never to be seen by anyone. All text sent to stdout and
> stderr are gone for good.
>
> Solution:
> Always always always always start with the --verbose flag. Now you will
> see every bit of output -- including Felix errors, services that
> blow-up, xml parsing errors, etc. It automatically shows all log
> messages. If you hit ^C the server is stopped -- gracefully. The only
> time you don't want to use --verbose if for automated tests.
>
> There already is a script that you can run -- it is called "startserv"
> and it is in the same directory as asadmin
>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe_at_glassfish.dev.java.net
> For additional commands, e-mail: dev-help_at_glassfish.dev.java.net
>