Rstudio hanging when using R installed in network drive

[mattybenwah]

Hi R Studio fans,

Looking for some advice on a set-up I have between home and the office. I have R installed on a network drive which allows me in theory to use it between a home laptop and an office PC. My method is that I have a local installation of R on the laptop which I then give a path to the network drive. This has worked ok in the past, but I have an issue where after I specify the path to the network drive library and then display the library paths via:

.libPaths(c(.libPaths(), "M:/R/win-library/3.6"))
.libPaths()

I get a very long hang (5-10 minutes) after executing the 2nd line despite it surely being a simple command. This extends to loading Bioconductor (a suite of packages within the network drive) which takes a similar unworkable amount of time. I'm wondering whether it would be related to this previous thread on issues with Rstudio with R installed on network drives though no solutions to improve speed seem to be given here:

RStudio and Network Drives

I'm getting a lot of hangs in RStudio because my user directory is on a network drive. I notice that session info (~/AppData\Local\RStudio-Desktop) is also on a network drive; is there any support for moving this folder? Simple evaluations / calculations hang for quite a while. Lookups for package documentation are also slow (every package is in Program Files, and I've removed the personal library from the search path). Any input would be much appreciated.

I have a good quality internet connection but am going through a VPN to get to the drive. Would appreciate any advice to diagnose what is happening here.

Cheers, Matt

[kevinushey]

What version of RStudio are you using? The latest release of RStudio (v1.2.1335) should in theory handle these sorts of issues better, but perhaps not.

When the R library paths are changed, RStudio will attempt to re-index packages installed in those libraries. I believe that can be slower when the library is on a networked filesystem. One workaround is to disable the Packages pane, e.g. with

Enable packages pane

[mattybenwah]

Thanks Kevin, no visible improvement with turning off the packages panel

Here's the output of libPaths() if it helps:

> .libPaths()
[1] "C:/Users/Matt/Documents/R/win-library/3.6" "C:/Program Files/R/R-3.6.1/library"       
[3] "M:/R/win-library/3.6"                     
Error: invalid version specification ‘NA’
In addition: Warning message:
In utils:::packageDescription(packageName, fields = "Version") :
  no package 'knitr' was found

Does seem like any work done downstream is at a reasonable speed aside from the annoyingly long start up time