[R] Snowfall: "cannont open connection " problem (repost in plain text)
Edmond.Ng at lshtm.ac.uk
Edmond.Ng at lshtm.ac.uk
Thu Jun 24 13:09:39 CEST 2010
<< Apologies for posting my question in HTML earlier. Reposted in plain text now.>>
Dear R-listers,
I have been using Snowfall (version 1.84) for parallel computing on a quad-PC running Windows 7 for a month or so without much problem. I started having problems runnnig R with Snowfall over our network when a new Novell client was installed for Windows 7. I experinenced network mapping problem with this new client. Our network colleague provided a fix which worked but a little unstable. So I decided to copy every file that I need from the networked drive to my local drive and ran R locally instead. I had no problem with R running locally for many long R jobs until this morning when Snowfall has stopped working completely. PLease note my R is installed locally.
Snowfall has been complaining about not being able to open a connection. I am not sure what it means as I am not a network person. Any advice will be very welcome. Many thanks in advance. See output and the error message below.
Best wishes,
Edmond
************** OUTPUT AND ERROR MESSAGE *******************
> # initialize Snowfall parameters
> sfInit(parallel=TRUE, cpus=2, type="SOCK")
snowfall 1.84 initialized (using snow 0.3-3): parallel execution on 2 CPUs.
> # perform parallel calculations
> rr <- list()
> date()
[1] "Thu Jun 24 09:17:56 2010"
> el.time <- system.time(
+
+ for (i in 1:(n.sims/everyout)) {
+ loop.start <- i*everyout -(everyout-1)
+ loop.end <- i*everyout
+ print(paste("loop.start = ",loop.start))
+ print(paste("loop.end = ",loop.end))
+
+ rr[[i]] <- sfClusterApplyLB(loop.start:loop.end,wrapper1,reps)
+ a<-rr[[i]]
+ save(a,file=paste("rr_",i,".RData",sep=""))
+ }
+ )
[1] "loop.start = 1"
[1] "loop.end = 10"
Error in checkForRemoteErrors(val) :
10 nodes produced errors; first error: cannot open the connection
Timing stopped at: 0 0 0.06
************** END OF OUTPUT AND ERROR MESSAGE *******************
More information about the R-help
mailing list