Thanks for a detailed answer, Robert.
Would have any workaround?
I was thinking of embedding a hardcoded 5s delayed redirects to a
gateway page into custom 500 error page.
Sincerely
Andre Rekhtine
IT Director, M.Sc., MCSE
Moveable Online Inc.
+1 (416) 532-5690
an...@moveable.com
-----Original Message-----
From: Robert Shubert [mailto:rshub...@tronics.com]
Sent: Thursday, July 28, 2011 12:38 PM
To: Witango-Talk@witango.com
Subject: RE: Witango-Talk: Witango Failover tune up
Andre,
The issue you specify where existing users don't failover is a known bug. Also,
Witango has never had the functionality to do stateful failover. In fact most
environments don't (by default) since it requires either a single point of
reference (such as a database) or a lot of overhead.
TeraScript will eventually have this feature, unfortunately, there are several
steps before we get there:
1. Correcting the load balancing and (stateless) failover capabilities of the
client <-> server protocol.
2. Developing a simple pooling system that can dynamically expand and shrink
the pool.
3. Adding persistent variable synchronization (sharing domain and application
scopes variables between pooled servers)
4. Adding user variable synchronization (sharing user scoped variables between
pooled servers)
We are also keen on correcting all underlying crashing issues so failover is
more of an administrative function than a safeguard.
All of this and lot more is on the list.
Robert
-----Original Message-----
From: Andre Rekhtine [mailto:and...@moveable.com]
Sent: Thursday, July 28, 2011 11:16 AM
To: Witango-Talk@witango.com
Subject: Witango-Talk: Witango Failover tune up
Gentlemen.
In the recent days we've formed a requirement for a stateful failover
among the application servers in the array.
In case of a Server crash, we want the customer to be transparently
carried over to another working daemon.
The way it works in our testlab, when a particular daemon crashes a
visitor is presented with the custom 500 error page. The problem is, that
customer doesn't get failed over to the next available Witango instance, while
every subsequent new visitor gets properly routed to the next working instance
by an ISAPI witango client plug-in.
Is there any built-in mechanism for a transparent failover of the
existing visitor sessions over to the next server in the array? Perhaps we're
missing some configuration settings?
Sincerely
Andre Rekhtine
IT Director, M.Sc., MCSE
Moveable Online Inc.
+1 (416) 532-5690
an...@moveable.com
----------------------------------------
To unsubscribe from this list, please send an email to lists...@witango.com
with "unsubscribe witango-talk" in the body.
----------------------------------------
To unsubscribe from this list, please send an email to lists...@witango.com
with "unsubscribe witango-talk" in the body.
----------------------------------------
To unsubscribe from this list, please send an email to lists...@witango.com
with "unsubscribe witango-talk" in the body.
