[
https://jira.jboss.org/jira/browse/JGRP-902?page=com.atlassian.jira.plugi...
]
Mitch Christensen commented on JGRP-902:
----------------------------------------
We also discussed the fact that the GR should be able to assume FD (PING) responsibility
as well,
{code:title=Chat snippet describing GR's implementation of FD as well as FD_SOCK}
[07:20] bela: #1 GR detects (abnormal) loss of a connection and sends a SUSPECT(X) to all
nodes in that group
[07:21] bela: (this is FD_SOCK functionality)
[07:21] MitchChristensen: I see...
[07:21] bela: #2 GR periodically sends a heartbeat to all GossipClients
[07:21] bela: ahem... RouterStubs (they've been merged)
[07:21] bela: (that's FD / FD_ALL functionality)
[07:22] MitchChristensen: I get it...so FD is completely transferred to the GR
[07:22] bela: With this scheme, you could remove FD and FD_SOCK from a tunnel.xml stack
[07:22] bela: Yes
[07:22] vlada_jgroups: hmmm, cool
{code}
GossipRouter: raise SUSPECT event when GR detects (abnormal) loss of
a connection from client
---------------------------------------------------------------------------------------------
Key: JGRP-902
URL:
https://jira.jboss.org/jira/browse/JGRP-902
Project: JGroups
Issue Type: Feature Request
Affects Versions: 2.6, 2.7
Reporter: Vladimir Blagojevic
Assignee: Vladimir Blagojevic
Fix For: 2.8
Currently we only support FD-heartbeat based failure detection algorithm for GR
configurations. While this works ok we do not have the same efficiency of failure
detection that comes with FD/FD_SOCK combo used in default UDP/TCP configurations. We can
bridge that gap by introducing FD_SOCK kind of functionality that is built in GR.
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
https://jira.jboss.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see:
http://www.atlassian.com/software/jira