Hi all,

I am currently evaluating a Wicket fail-over solution (two equal nodes with
a floating IP, probably using Linux Heartbeat) with Terracotta. Of course,
it would be desirable to switch between nodes without any impact on a user's
session. In order to provide this, the cluster must be able to handle
non-sticky sessions. Session clustering works like a breeze with Wicket and
Terracotta

I already spotted a problem with buffered redirects, that could be avoided
with a different rendering strategy
(IRequestCycleSettings.REDIRECT_TO_RENDER). However, this problem would only
occur if a node goes down just when the client is redirecting - not very
likely and probably acceptable as switching over might take a few seconds
anyway.

Now my question: What other problems could appear when I only use session
clustering for fail-over purposes? Can I stay with the default rendering
strategy (IRequestCycleSettings.REDIRECT_TO_BUFFER) without clustering the
buffer if I accept missing feedback messages in the above scenario?

Any further comments on my setup will be appreciated as well!

(If answers are woth it, I'll create a wiki page out of this thread)

Best Regards, Stefan


-----
-------
Stefan Fußenegger
http://talk-on-tech.blogspot.com // looking for a nicer domain ;)
-- 
View this message in context: 
http://www.nabble.com/Wicket-fail-over-with-Terracotta---sticky-or-non-sticky-sessions--tp16467830p16467830.html
Sent from the Wicket - User mailing list archive at Nabble.com.


---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to