Regeristering with Reggie and "Read timed out"

Hi all,

We had an incident in production this past weekend we had not seen
before and I was wondering if anyone else has experienced this or might
offer some insight.

Following our traditional procedure for rolling out new patches, we
brought down some services, patched the code and then started them up
again at which time they attempted to register with the LUSs.  Normally
this process is pretty automatic and the services are immediately
available to the clients.

However this time during registration the logs were filled with "Read
timed out" exceptions and the services were unable to register.  The
rest of the service layer continued to function as normal but we could
not get these new services to register.  In a drastic measure we bounced
all of our LUSs (4) and then started the services experiencing the "Read
timed out" issues.  Within minutes the whole services layer was back up
and running having only bounced the LUSs and the initial set of services
we were patching.

I read through some of the Jini docs and it looks like Reggie waits some
amount of time for a client to handshake during registration before
giving up on it (admittedly I'm not completely clear on this).  Why
would bouncing the LUSs fix this issue if it was a client problem?

We have a thread dump of one of the LUSs during this issue if it's of
use to anyone.  We never recalled having seen this issue before.
Fortunately this happened during a slow period of traffic so bouncing
the LUSs didn't have much impact but should this be our SOP?  Is this
something anyone else has experienced?

thanks, brian

--------------------------------------------------------------------------
Getting Started:     http://www.jini.org/wiki/Category:Getting_Started
Community Web Site:  http://jini.org
jini-users Archive:  http://archives.java.sun.com/archives/jini-users.html
Unsubscribing:       email "signoff JINI-USERS"  to [email protected]