search

mbraceproject / MBrace.StarterKit

A collection of demos and tutorials for MBrace
http://mbrace.io
57 stars 34 forks source link

If worker die, process status never been reset #21

Closed soloman817 closed 9 years ago

soloman817 commented 9 years ago

Hi guys,

I have 2 local workers, and one worker is slow to download assembly, so I just close it, and want to start a new local worker. But looks like the process is always "Running", so my last cloud call is always blocked. I think in this case, return exception is better than just blocking forever.

see the screenshot: 2015-05-01_1156

krontogiannis commented 9 years ago

Hello, is your code using CloudFlow or Parallel/StartChild with WorkerRef affinity?

soloman817 commented 9 years ago

@krontogiannis , I think I used CloudFlow

krontogiannis commented 9 years ago

Ok, so in cases with deterministic scheduling like CloudFlow or Parallel/etc with IWorkerRef, fault tolerance is incomplete (when the particular node doesn't restart or recover). I will close this issue, and post any updates here : https://github.com/mbraceproject/MBrace.Azure/issues/77