-
Suggestion
-
Resolution: Obsolete
-
None
-
None
-
« Hide
Microsoft Windows Server 2003 Standard Edition Service Pack 1
Intel Xeon CPU 2 GHz / 2 GHz, 2 GB of RAM
On the issue regarding certain IP addresses not working for the clustering, we talked to our Cisco support group and they came back with the following:
'It looks like for the non-working address (225.137.66.56) the last IGMP message was CHANGE_TO_EXCLUDE_MODE which will block the IGMP traffic for the group. For the working address 23.0.0.5 the last one was CHANGE_TO_INCLUDE_MODE which forward traffic to the mcast group. Could you please check with the multicast test software provider ?'
Any idea why one address would use TYPE - CHANGE_TO_EXCLUDE_MODE?
[CONFSERVER-10330] Certain Multicast IP addresses sending CHANGE_TO_EXCLUDE_MODE causing IGMP traffic to be blocked
Christopher Owen [Atlassian]
added a comment - Hi Justin,
That makes more sense. In that case, we really can't explain the behaviour. There is no reason why messages sent to one multicast IP address should be any different to another valid multicast IP address. Given that we've only had the solitary report of behaviour like this from any of our clustered customers and all of our testing internally hasn't exhibited this behaviour it is unlikely that we are going to be able to spend much time investigating this further.
If you'd like some more follow up I could send this report to Tangosol (the providers of Confluence's basic cluster functionality) to see if they have any further commentary on this issue. At the moment it would appear to be something perculiar to the host OS or network.
Chris
Christopher Owen [Atlassian]
added a comment - Hi Justin,
That makes more sense. In that case, we really can't explain the behaviour. There is no reason why messages sent to one multicast IP address should be any different to another valid multicast IP address. Given that we've only had the solitary report of behaviour like this from any of our clustered customers and all of our testing internally hasn't exhibited this behaviour it is unlikely that we are going to be able to spend much time investigating this further.
If you'd like some more follow up I could send this report to Tangosol (the providers of Confluence's basic cluster functionality) to see if they have any further commentary on this issue. At the moment it would appear to be something perculiar to the host OS or network.
Chris 
Justin Engelking
added a comment - I believe that email might have been mistyped. The address I believe we are currently using in our production environment is 237.0.0.5. I was able to get this to work by manually editing the confluence.cfg.xml file on the server.
Justin Engelking
added a comment - I believe that email might have been mistyped. The address I believe we are currently using in our production environment is 237.0.0.5. I was able to get this to work by manually editing the confluence.cfg.xml file on the server. Christopher Owen [Atlassian]
added a comment - Hi Justin,
I've just been reviewing this issue again. Is the reported working IP address really 23.0.0.5? That address isn't actually a valid multicast IP address. How did you actually configure Confluence to use that address? I find it very surprising that Confuence actually works using that address.
Chris
Christopher Owen [Atlassian]
added a comment - Hi Justin,
I've just been reviewing this issue again. Is the reported working IP address really 23.0.0.5? That address isn't actually a valid multicast IP address. How did you actually configure Confluence to use that address? I find it very surprising that Confuence actually works using that address.
Chris Justin Engelking
added a comment - Can I have an update on this issue? We have Cisco support waiting on the sesults of this as well. Thanks.
Justin Engelking
added a comment - Can I have an update on this issue? We have Cisco support waiting on the sesults of this as well. Thanks. Christopher Owen [Atlassian]
added a comment - I'm not entirely sure why this would be happening - it's certainly nothing we've seen before and could potentially be OS specific. We'll let you know as soon as we are able to research this some more
Christopher Owen [Atlassian]
added a comment - I'm not entirely sure why this would be happening - it's certainly nothing we've seen before and could potentially be OS specific. We'll let you know as soon as we are able to research this some more Justin Engelking
added a comment - Just checking on the status of this. Is this a verified bug? Or is there some reason why the TYPE is CHANGE_TO_EXCLUDE_MODE on certain IP addresses?
Justin Engelking
added a comment - Just checking on the status of this. Is this a verified bug? Or is there some reason why the TYPE is CHANGE_TO_EXCLUDE_MODE on certain IP addresses?
Please do send this on to Tangosol. I would like to be updated on the progress of this. Thanks.