[
https://issues.apache.org/jira/browse/GUACAMOLE-1022?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17104561#comment-17104561
]
Jerome Charaoui commented on GUACAMOLE-1022:
--------------------------------------------
I've been testing session affinity and I also think there could be improvements.
For example, if a connection becomes unavailable because the connexion weight
is changed to "-1", the Load Balanced Group will fail for that user until they
log out. The same happens if the client was disconnected, and in the mean time
another user logged into the target computer.
In both of these cases the error message displayed is "This connection is
currently in use, and concurrent access to this connection is not allowed.
Please try again later." If the user doesn't care about their previous
connection, it's not clear at all that what they need to do is log out and log
back in.
Perhaps a more acceptable approach to this problem would be to implement
*optional* failover. If for whatever reason the client is unable to reestablish
the connection using the affinity parameter, instead of failing hard it could
display a more informative error message which includes the option to attempt
to reconnect to the Load Balanced Group while ignoring affinity parameters. For
example:
"We were unable to reconnect you to your previous connection. Would you like to
attempt to reconnect using a new connection?"
> Allow load balancing groups to fail to next connection
> ------------------------------------------------------
>
> Key: GUACAMOLE-1022
> URL: https://issues.apache.org/jira/browse/GUACAMOLE-1022
> Project: Guacamole
> Issue Type: Improvement
> Components: guacamole
> Reporter: Jason Keltz
> Priority: Minor
>
> I have 78 hosts in a balanced connection group. Let's say that the first
> host is down. Now, when the user click on the connection group expecting to
> get a connection, they instead get a message that Guacamole can't connect.
> If it can't connect to the first one, it should try the next one, then the
> next one until it finds one that works.
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
