Re: [Mpi3-ft] Process failure document
Folks, I've added some comments to the document, which you'll find appended to this message. Agree with Rich that getting the model for colective comms right has high priority, so it's OK to focus on this one. Questions I'd have to the group are - is our model symmetric - that is if 2n processes are split in the middle into two groups of n processes, will each group see the same errors and be able to recover? - supose processes A and B. A fails after a while, B notices and starts a repair with restore processes; in which state will a process A' be whrn it joins B after the repair? Will A' need to do any special repair call itself? Cheerio Hans-Christian Hoppe Principal Engineer Intel GmbH Phone: +49-2232-2090-11 Hermuelheimer Strasse 8a Fax: +49-2232-2090-29 50321 Bruehl, Germany ________________________________ From: [email protected] [mailto:[email protected]] On Behalf Of Richard Graham Sent: Dienstag, 4. November 2008 22:34 To: MPI 3.0 Fault Tolerance and Dynamic Process Control working Group Subject: [Mpi3-ft] Process failure document I have captured a lot of what we have discussed about process fault-tolerance, and filled in more missing gaps to help move us a long a bit faster in our discussions. Please take a look at the document before the call tomorrow. I would like to pick up discussing what to do when collective communications fail. There are still details missing that need to be added. No API's at this stage, just the "model". I ran this past 3 different application groups today - this seems to be along the lines of what they are looking for, and they had some very useful comments... Rich --------------------------------------------------------------------- Intel GmbH Dornacher Strasse 1 85622 Feldkirchen/Muenchen Germany Sitz der Gesellschaft: Feldkirchen bei Muenchen Geschaeftsfuehrer: Douglas Lusk, Peter Gleissner, Hannes Schwaderer Registergericht: Muenchen HRB 47456 Ust.-IdNr. VAT Registration No.: DE129385895 Citibank Frankfurt (BLZ 502 109 00) 600119052 This e-mail and any attachments may contain confidential material for the sole use of the intended recipient(s). Any review or distribution by others is strictly prohibited. If you are not the intended recipient, please contact the sender and delete all copies.
participants (1)
-
Hoppe, Hans-Christian