For the benefit of the archive:

On 10/09/2010, at 9:32 PM, Paul Hoadley wrote:

> Turns out I've broken something since this (HelloWorld) was working in 
> JavaMonitor yesterday (I swear!).  Looks like wotaskd starts up fine, but 
> JavaMonitor shows this:
> 
> [2010-9-10 11:49:36 UTC] <main> Creating LifebeatThread now with: JavaMonitor 
> 8083 ip-10-166-105-197/10.166.105.197 1085 30000
> [2010-9-10 11:49:36 UTC] <main> WOLocalRootDirectory set to non-default: 
> /opt/Local
> [2010-9-10 11:49:36 UTC] <Thread-1> <_LifebeatThread> Exception creating 
> datagramSocket: java.net.BindException: Cannot assign requested address
> 
> Is that telling me JavaMonitor can't talk to wotaskd?

This was the real problem.  The 'stalling before "Waiting for requests..."' 
thing was a red herring, maybe I didn't have logging turned up high enough to 
see that, who knows.  In any case, the app _was_ starting up from the command 
line, but the communication between wotaskd and JavaMonitor had become 
completely hosed, so it never got close to starting from JavaMonitor.  This is 
the trap for young players, and it's probably fairly specific to Amazon's EC2: 
when you stop an EBS-backed instance and re-start it, it seems you get a new 
public-facing hostname, _but_ the running image itself doesn't seem to realise 
it.  Both 'hostname' and /etc/hostname report the old name.  This seemed 
terminal to wotaskd/JavaMonitor, and setting -WOHost didn't seem to help.

In any case, everything is working fine on a brand new instance.  Nothing to 
see here.


-- 
Paul.

http://logicsquad.net/


 _______________________________________________
Do not post admin requests to the list. They will be ignored.
Webobjects-deploy mailing list      ([email protected])
Help/Unsubscribe/Update your Subscription:
http://lists.apple.com/mailman/options/webobjects-deploy/archive%40mail-archive.com

This email sent to [email protected]

Reply via email to