Skip to content

R crashes when using huge data sets with character string variables

7 messages · Arne Henningsen, Ben Bolker, brodie gaslam +3 more

Original
Arne Henningsen
# 
When working with a huge data set with character string variables, I
experienced that various commands let R crash. When I run R in a
Linux/bash console, R terminates with the message "Killed". When I use
RStudio, I get the message "R Session Aborted. R encountered a fatal
error. The session was terminated. Start New Session". If an object in
the R workspace needs too much memory, I would expect that R would not
crash but issue an error message "Error: cannot allocate vector of
size ...".  A minimal reproducible example (at least on my computer)
is:

nObs <- 1e9

date <- paste( round( runif( nObs, 1981, 2015 ) ), round( runif( nObs,
1, 12 ) ), round( runif( nObs, 1, 31 ) ), sep = "-" )

Is this a bug or a feature of R?

Some information about my R version, OS, etc:

R> sessionInfo()
R version 4.0.3 (2020-10-10)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: Ubuntu 20.04.1 LTS

Matrix products: default
BLAS:   /usr/lib/x86_64-linux-gnu/blas/libblas.so.3.9.0
LAPACK: /usr/lib/x86_64-linux-gnu/lapack/liblapack.so.3.9.0

locale:
[1] LC_CTYPE=en_DK.UTF-8       LC_NUMERIC=C
[3] LC_TIME=en_DK.UTF-8        LC_COLLATE=en_DK.UTF-8
[5] LC_MONETARY=en_DK.UTF-8    LC_MESSAGES=en_DK.UTF-8
[7] LC_PAPER=en_DK.UTF-8       LC_NAME=C
[9] LC_ADDRESS=C               LC_TELEPHONE=C
[11] LC_MEASUREMENT=en_DK.UTF-8 LC_IDENTIFICATION=C

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base

loaded via a namespace (and not attached):
[1] compiler_4.0.3

/Arne
Ben Bolker
# 
On Windows you can use memory.limit.

https://stackoverflow.com/questions/12582793/limiting-memory-usage-in-r-under-linux

    Not sure how much that helps.
On 12/12/20 6:19 PM, Arne Henningsen wrote:
brodie gaslam
# 
On OS X I see:

??? > nObs <- 1e9
??? >? date <- paste( round( runif( nObs, 1981, 2015 ) ), round( runif( nObs,1, 12 ) ), round( runif( nObs, 1, 31 ) ), sep = "-" )
??? Error: vector memory exhausted (limit reached?)
??? > sessionInfo()
??? R version 4.0.3 (2020-10-10)
??? Platform: x86_64-apple-darwin17.0 (64-bit)
??? Running under: macOS Catalina 10.15.7

Which is what I would expect.? I don't doubt the error you've seen, just
providing a data point for whoever ends up looking into this further.

Best,

Brodie.
iuke-tier@ey m@iii@g oii uiow@@edu
# 
If R is receiving a kill signal there is nothing it can do about it.

I am guessing you are running into a memory over-commit issue in your OS.
https://en.wikipedia.org/wiki/Memory_overcommitment
https://engineering.pivotal.io/post/virtual_memory_settings_in_linux_-_the_problem_with_overcommit/

If you have to run this close to your physical memory limits you might
try using your shell's facility (ulimit for bash, limit for some
others) to limit process memory/virtual memory use to your available
physical memory. You can also try setting the R_MAX_VSIZE environment
variable mentioned in ?Memory; that only affects the R heap, not
malloc() done elsewhere.

Best,

luke
On Sat, 12 Dec 2020, Arne Henningsen wrote:

            

      
      
      
    

        

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      Dirk Eddelbuettel
    

    

      
      #
    

  


  

      
On 12 December 2020 at 21:26, luke-tierney at uiowa.edu wrote:

        
| If R is receiving a kill signal there is nothing it can do about it.
| 
| I am guessing you are running into a memory over-commit issue in your OS.
| https://en.wikipedia.org/wiki/Memory_overcommitment
| https://engineering.pivotal.io/post/virtual_memory_settings_in_linux_-_the_problem_with_overcommit/
| 
| If you have to run this close to your physical memory limits you might
| try using your shell's facility (ulimit for bash, limit for some
| others) to limit process memory/virtual memory use to your available
| physical memory. You can also try setting the R_MAX_VSIZE environment
| variable mentioned in ?Memory; that only affects the R heap, not
| malloc() done elsewhere.

Similarly, as it is Linux, you could (easily) add virtual memory via a
swapfile (see 'man 8 swapfile' and 'man 8 swapon').  But even then, I expect
this to be slow -- 1e9 is a lot.

I have 32gb and ample swap (which is rarely used, but a safety net). When I
use your code with nObs <- 1e8 it ends up with about 6gb which poses poses no
problem, but already takes 3 1/2 minutes:

            

      
      
      
    

        
user  system elapsed 
203.723   1.779 205.528

            

      
      
      
    

        
You may want to play with the nObs value to see exactly where it breaks on
your box.

Dirk

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      Iñaki Ucar
    

    

      
      #
    

  


  

      
On Sun, 13 Dec 2020 at 04:27, <luke-tierney at uiowa.edu> wrote:

            

      
      
      
    

        
Correct. And in particular, this is most probably the earlyoom [1]
service in action, which, I believe, is installed and enabled by
default in Ubuntu 20.04. It is a simple daemon that monitors memory,
and when some conditions are reached (e.g., the system is about to
start swapping), it looks for offending processes and kills them.

[1] https://github.com/rfjakob/earlyoom

I?aki

            

      
      
      
    

        

      
      
    

        

  
    

      
      
    

  


  


      

    

    
1 day later

    

      
      

        


  

        

      Arne Henningsen
    

    

      
      #
    

  


  

      
Dear all

Thanks a lot for your very helpful explanations and suggestions. I
have increased the size of my computer's "swapfile" and this solved my
problem, i.e., R no longer crashes when I work with character string
variables in my large data set (probably until I work with an even
larger data set).

Best wishes,
Arne

        
On Sun, 13 Dec 2020 at 11:17, I?aki Ucar <iucar at fedoraproject.org> wrote: