Meeting recording
Unfortunately started after the first talk. Will do better next time.
Participants:
Alessandra Forti, Alexey Sedov, Alison
McCrea, Andrej Filipcic, Andrew Lahiff, Andrew
McNab, Antonio Maria Perez Calero Yzquierdo, Carles, Daniel Peter Traynor, Gareth Roy, Jeff Templon, Jerome Belleman, Jose Flix Molina, Jose Hernandez Calama, Manfred Alef, Peter Love, Raul (brunel), Recorder, Sam Skipsey, Simone Campana, Stefano Dal Pra, Thomas Hartmann
Chat log
Carles: (18/02/2014 13:52)
I'm sorry but know I don't hear anything
I am restarting vidyo
Jeff Templon: (01:52 PM)
ok ... i do hear ...
Sam Skipsey: (01:53 PM)
(I note that the Glue2.0 schema doesn't seem to use SMPSize anymore? IIRC, Glue2.0 is supposed to let you publish information for each "set" of identical resources, though.)
(So, you just publish CPUMultiplicity values for each subset of the cluster)
Jeff Templon: (01:54 PM)
good question. the link i just sent only deals with glue v1.3
Sam Skipsey: (01:55 PM)
For Glue1.3, are you not supposed to publish one
SubCluster for each homogeneous set?
So, if you have N types of host, you need N
SubClusters, each with the correct SMPSize set, ideally
Andrew
McNab: (02:13 PM)
the other thing is to have elastic jobs that look at how much time they have left and keep running till then
Sam Skipsey: (02:23 PM)
Sorry, Andrew, you're quite right: we did diverge from that conversation before I got a chance to mention your comment
--
AntonioPerezCalero - 17 Feb 2014