1

I have 2 workstations with different processors, both on Windows 10 Pro. I want to create cluster on both computers. 

primary <- '220.168.1.10'
machineAddresses <- list(list(host=primary, ncore=maxThreads), list(host='220.168.1.11', ncore=maxThreads));

spec <- lapply(machineAddresses,
               function(machine) 
           {
                 rep(list(list(host=machine$host, user=machine$user)), machine$ncore)
               })

spec <- unlist(spec,recursive=FALSE)

internalCluster <- makeCluster(type='PSOCK', master=primary, spec=spec, homogeneous=FALSE);

But code execution holds on the last string of this code. I use this source: http://www.win-vector.com/blog/2016/01/running-r-jobs-quickly-on-many-machines/ So, what does it need to do for running cluster on different machines?

Thank you.

Dmitriy
  • 847
  • 17
  • 39
  • This is the classic "makeCluster" hangs problem, but in this case it's worse because ssh is used to start the workers. Is ssh working properly between your machines? Traditionally that's been a big problem on Windows. – Steve Weston Feb 04 '17 at 13:07
  • @Steve Weston, no, I didn't configure ssh. I installed OpenSSH, but didn't do anything for configure this. Is there are article about correct ssh configuration for cluster making in Windows? Thnx – Dmitriy Feb 04 '17 at 15:15
  • I've heard that there is now a native ssh distribution available from Microsoft, but I've never used it. I suggest that you search for information about that. Make sure you install an ssh client on the master machine and ssh servers on all of the worker machines. That's what is needed by `makeCluster`. – Steve Weston Feb 04 '17 at 15:44
  • See http://stackoverflow.com/a/17976175/2109128 and http://stackoverflow.com/a/17551327/2109128 for more information on this issue. – Steve Weston Feb 04 '17 at 15:46
  • @Steve Weston, thank you much! – Dmitriy Feb 04 '17 at 15:59

0 Answers0