[Spread-users] Spread 4.0 daemon "goes to sleep" on XP network disconnect

[Yair Amir]

Hi Steve,

In my opinion, Spread can be made to handle this situation, although
it is not a 5 minutes fix. If my memory serves me well, the only meaningful
place where a network is relied upon in a case on a singleton, is
in passing the token from the local daemon to the local daemon.
A potential strategy would be to add code that in the case of
a daemon that wants to send the token to itself, it will somehow
bypass the network (it can also send it to localhost but this seems
ugly and breaks Spread clean separation). Maybe Jonathan can
have a better insight.

This is not the only relevant such unicasts - spmonitor is
also using these, but I think this strategy will solve the core problem.

If you really need this, perhaps the quick ugly fix is what you need (checking
on token sends if they are destined to this daemon and if so just send them to
127.0.0.1 instead).

Cheers,

	:) Yair.

Steve Duff wrote:
> Hi,
> 
> Thanks for taking the time to reply.
> 
> You are correct that spsend cannot send to the local IP address when the
> network card has been disconnected - it can send to 127.0.0.1, but
> spread is not using this address for "send to self".
> 
> It appears that unlike other OS's, XP deletes (or makes inaccessible)
> the entire interface configuration when it detects the failure of the
> network connection, so can't possibly route traffic to the local IP
> address. 
> 
> According to one of my colleagues, previous versions of Windows, e.g.
> Win2k, do not do this. I'm sure that Microsoft has a good reason for
> introducing this behaviour in XP, but I can't think what it might be.
> 
> Thanks again for your help.
> 
> Steve
> 
> -----Original Message-----
> From: Yair Amir [mailto:yairamir at cs.jhu.edu] 
> Sent: 08 August 2007 15:36
> To: Steve Duff
> Cc: spread-users at lists.spread.org
> Subject: Re: [Spread-users] Spread 4.0 daemon "goes to sleep" on XP
> network disconnect
> 
> Hi,
> 
> It looks to me that in your case, the operating system does not pass
> messages
> from the machine to itself when you unplug the cable. This means that
> Spread cannot send a unicast to itself on the local node. You can verify
> this by
> using the two programs spsend and sprecv provided with Spread (you may
> need to build them - "make spsend", etc.)
> 
> Run both of them on the singleton machine before you unplug to verify
> that unicasts from spsend to sprecv using the IP address from the
> configuration file will be received. Then, unplug and try again.
> Then plug and try again. If my observation is correct,
> spsend will not be able to pass messages to sprecv on your system
> locally
> on the same machine when the cable is unplugged. It may be that it
> even drops the whole interface if the cable is not plugged.
> 
> Let us know if this is the case.
> 
> Cheers,
> 
> 	:) Yair.
> 
> 
> **********************************************************************
> 
> SunGard Vivista Limited, Marshfield, Chippenham, Wiltshire  SN14 8SR
> Telephone: 08456 041999,  Fax: 08456 052999
> 
> Registered Office:  33 St Mary Axe, London EC3A 8AA.  Registered in England No. 1593831 VAT Reg No. GB 810 9546 34
> 
> **********************************************************************
> 
> This email and any files transmitted with it are confidential and
> intended solely for the use of the individual or entity to whom they
> are addressed. If you have received this email in error please notify
> the system manager.
> 
> This footnote also confirms that this email message has been swept by
> MIMEsweeper for the presence of computer viruses.
> 
> www.mimesweeper.com
> **********************************************************************
> 
> 
> This message has been checked for all known viruses on behalf of SunGard Vivista by MessageLabs. 
> 
> http://www.messagelabs.com or Email: mailsweeper.info at vivista.sungard.com
> 
> For further information http://www.sungard.com/vivista
> 
> _______________________________________________
> Spread-users mailing list
> Spread-users at lists.spread.org
> http://lists.spread.org/mailman/listinfo/spread-users
> 
>