[erlang-questions] Resisting "noconnection" / Remote termination of nodes

Kannan <>
Thu Apr 26 12:34:39 CEST 2012


Hi Olivier,

In distributed computing, we see the collective effort of individual -- but
not independent -- nodes. Communicating nodes, at any point in time, need
to be connected and should have the knowledge of the other, including the
availability status. The OAM node, or leader, will have the total picture
of all the nodes.

You have to look at the bigger picture of the whole system, when it comes
to stopping a distributed system. The following steps are generally taken
to tear down a system.

* Signal all the nodes of a forthcoming shutdown request
* Stop accepting new requests
* Finish servicing the accepted requests
* Do the stock taking and clean-up
* Check the readiness of all the nodes for shutdown
* Then call shutdown on all the nodes.

Here, I am looking from top into the system. The shutdown process will be
generally coordinated by a single process.

An ad-hock shutdown is more scary to me in a production environment.

Kind Regards,
Kannan.




On Tue, Apr 24, 2012 at 5:22 PM, Olivier BOUDEVILLE <
> wrote:

>
> Hi,
>
> For a more controlled overall termination of a distributed application, I
> try to shutdown synchronously a series of nodes, as properly and as in
> parallel as possible, in a non-OTP program. I imagine that using '[
> rpc:cast( N, erlang, halt, [] ) || N <- MyTargetNodes ]' and then waiting
> for them to be terminated is the best approach for that.
>
> As I want now these terminations to be synchronous (i.e. I want my
> terminate function to return only when all nodes are down for sure), I used
> to rely on checking their termination using net_adm:ping/1 (waiting for
> pong to become pang), but kept on getting (systematically) 'noconnection'
> errors (exceptions?), which do not seem to be catchable (at least not with
> a 'try .. catch T:E ->.. end' clause). This happens as soon as there is at
> least one node (which happens to be on the same host - of course it is not
> the local node from which that rpc:cast is triggered) to halt.
>
> I switched to looping on 'lists:member( Nodename, nodes() )' instead of
> ping (in both case with a proper waiting between checks), but I still get
> 'noconnection' errors. It looks like 'noconnection' is VM-level? As
> expected, commenting-out the rpc:cast/3 never leads to 'noconnection'.
>
> I feel I would need something like net_kernel:unconnect_node/1.
>
> My question now: how to deal gracefully with such a synchronous node
> shutdown and to resist to the (intended) loss of node(s)?
>
> Thanks in advance for any hint!
> Best regards,
>
> Olivier.
> ---------------------------
> Olivier Boudeville
>
> EDF R&D : 1, avenue du Général de Gaulle, 92140 Clamart, France
> Département SINETICS, groupe ASICS (I2A), bureau B-226
> Office : +33 1 47 65 59 58 / Mobile : +33 6 16 83 37 22 / Fax : +33 1 47
> 65 27 13
>
>
> Ce message et toutes les pièces jointes (ci-après le 'Message') sont
> établis à l'intention exclusive des destinataires et les informations qui y
> figurent sont strictement confidentielles. Toute utilisation de ce Message
> non conforme à sa destination, toute diffusion ou toute publication totale
> ou partielle, est interdite sauf autorisation expresse.
>
> Si vous n'êtes pas le destinataire de ce Message, il vous est interdit de
> le copier, de le faire suivre, de le divulguer ou d'en utiliser tout ou
> partie. Si vous avez reçu ce Message par erreur, merci de le supprimer de
> votre système, ainsi que toutes ses copies, et de n'en garder aucune trace
> sur quelque support que ce soit. Nous vous remercions également d'en
> avertir immédiatement l'expéditeur par retour du message.
>
> Il est impossible de garantir que les communications par messagerie
> électronique arrivent en temps utile, sont sécurisées ou dénuées de toute
> erreur ou virus.
> ____________________________________________________
>
> This message and any attachments (the 'Message') are intended solely for
> the addressees. The information contained in this Message is confidential.
> Any use of information contained in this Message not in accord with its
> purpose, any dissemination or disclosure, either whole or partial, is
> prohibited except formal approval.
>
> If you are not the addressee, you may not copy, forward, disclose or use
> any part of it. If you have received this message in error, please delete
> it and all copies from your system and notify the sender immediately by
> return message.
>
> E-mail communication cannot be guaranteed to be timely secure, error or
> virus-free.
>
> _______________________________________________
> erlang-questions mailing list
> 
> http://erlang.org/mailman/listinfo/erlang-questions
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://erlang.org/pipermail/erlang-questions/attachments/20120426/632002ff/attachment.html>


More information about the erlang-questions mailing list