[Spread-users] How to detect if a Spread node fails

[John Schultz]

You don't get a leave message but you do get a caused by network membership change message.  I'm saying you have to remember what the previous membership was, inspect the new membership and notice that the *#able guys are no longer there.

Cheers!

-----
John Lane Schultz
Spread Concepts LLC
Phn: 301 830 8100
Cell: 443 838 2200

On Jul 13, 2011, at 1:03 PM, Andrew Holt wrote:

Hi,

The interesting thing is I don’t get an leave messages.

To use my example I kill spread on node able and I get join messages for 


#bill#baker
#who#baker

I presume this is caused by the spread network reforming around the failed node.

If I got leave messages for:

fred#able
#joe#able

That would be fine.

Regards,
Andrew

On 13 Jul 2011, at 17:20, John Schultz wrote:

> Have at least one process per daemon join the groups in which you are interested in tracking.  Track the membership of the group.  If you see a caused by network where all the *#able members leave, then its a good bet that daemon has either crashed or partitioned away.
> 
> If that's not good enough then you might be able to administratively, through code changes, get a report of the actual daemon membership which can tell you if able is in the daemon membership or not.
> 
> Cheers!
> 
> -----
> John Lane Schultz
> Spread Concepts LLC
> Phn: 301 830 8100
> Cell: 443 838 2200
> 
> On Jul 13, 2011, at 11:27 AM, Andrew Holt wrote:
> 
> Hi,
> 
> I have two machines each running the spread server, and each setup so they know about each other.  Let’s call them able & baker.
> 
> I have a number of clients programs, in the same group on each
> 
> So: 
> 
> #fred#able
> #joe#able
> #bill#baker
> #who#baker
> 
> If spread dies, is killed or a machine dies how can I detect that the node has gone ‘offline’ ?
> 
> baker receives a message of the type Is_caused_network_mess for each group.  What I need is to know the source is 'able’ 
> 
> Any hints ?
> 
> Thanks,
> Andrew
> 
> =============================
> Andrew Holt
> 
> Email: andrew.holt at 4asolutions.co.uk
> 
> De Omnibus Dubitandum
> =============================
> 
> 
> 
> 
> _______________________________________________
> Spread-users mailing list
> Spread-users at lists.spread.org
> http://lists.spread.org/mailman/listinfo/spread-users
> 

=============================
Andrew Holt

Email: andrew.holt at 4asolutions.co.uk

De Omnibus Dubitandum
=============================




-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 3805 bytes
Desc: not available
Url : http://lists.spread.org/pipermail/spread-users/attachments/20110713/27e55980/attachment.bin