简体   繁体   English

在Hortonworks Sandbox hadoop 2.3上安装R

[英]Installing R on Hortonworks Sandbox hadoop 2.3

I'm experiencing some troubles when I try to install R dependencies with the command : 尝试使用命令安装R依赖项时遇到一些麻烦:

install.packages(c(“Rcpp”,”RJSONIO”,”bitops”,”digest”,”functional”,”itertools”,”reshape2″,”string    ”,”plyr”,”caTools”),repos=’http://cran.revolutionanalytics.com’)

packages are downloaded but then all I get are the messages 软件包已下载,但我得到的只是消息

/usr/lib64/R/bin/R: line 8: uname: command not found /usr/lib64/R/bin/R: line 143: exec: sh: not found / usr / lib64 / R / bin / R:第8行:uname:找不到命令/ usr / lib64 / R / bin / R:第143行:exec:sh:找不到

repeated as many times as the number of packages I am trying to install. 重复我尝试安装的软件包数量的次数。

I am using R version 3.2.2, rstudio-server 0.99.489 in Hortonworks Sandbox with hadoop 2.3 (Oracle Virtualbox). 我在带有Hadoop 2.3(Oracle Virtualbox)的Hortonworks Sandbox中使用R版本3.2.2,rstudio-server 0.99.489。

Do you have any suggestions? 你有什么建议吗?

Thanks for your help. 谢谢你的帮助。

I know this question was asked 7 months ago. 我知道这个问题是在7个月前提出的。 But I just had the same problem stefanod did and I Googled left and right without finding a solution. 但是我也遇到了stefanod遇到的同样问题,我左右搜索Google,但没有找到解决方案。 Then for absolutely no reasonable reason it just worked. 然后,出于绝对没有合理的原因,它只是起作用了。 Here is what I have tried and what eventually solves my problem: 这是我尝试过的方法,最终解决了我的问题:

So I use R 3.2.2 , rstudio-server 0.98.994 in Hortonworks Hadoop Sandbox HDP 2.4 in VMWare. 所以我在VMWare的Hortonworks Hadoop沙盒HDP 2.4中使用R 3.2.2rstudio-server 0.98.994 I logged in the virtual machine using the Shell Web Client Method . 我使用Shell Web Client Method登录了虚拟机。 The entire process was done with root privilege. 整个过程都具有root特权。 I followed this instruction: https://jsolderitsch.wordpress.com/hortonworks-sandbox-r-and-rstudio-install/ except that I used updated versions of R and RStudio Server . 我遵循了以下说明: https : //jsolderitsch.wordpress.com/hortonworks-sandbox-r-and-rstudio-install/,除了我使用RRStudio Server更新版本。 Everything worked fine until I invoked R in the shell and attempted 一切正常,直到我在外壳中调用R并尝试

install.packages('foo.bar', repos='http://cran.revolutionanalytics.com')

Then I had exactly the same error messages as in this Q. 然后,我得到了与该问题完全相同的错误消息。

I checked my $PATH and used whereis COMMANDNAME to check if I had the uname command at the right place. 我检查了$PATH并使用whereis COMMANDNAME检查在正确的位置是否有uname命令。 Seems I do. 好像我知道

I then quit R . 然后我退出R Inside CentOS I used wget to manually download the .tar.gz files of the packages and their dependencies. CentOS内部,我使用wget来手动下载软件包及其相关性的.tar.gz文件。 I then tried to install those packages manually: 然后,我尝试手动安装这些软件包:

[root@sandbox ~]# sudo R CMD INSTALL foo.bar.tar.gz                                                                                    
* installing to library ‘/usr/lib64/R/library’                                                                                              
* installing *source* package ‘foo.bar’ ...                                                                                                  
** package ‘foo.bar’ successfully unpacked and MD5 sums checked                                                                              
** libs                                                                                                                                     
sh: make: command not found                                                                                                                 
ERROR: compilation failed for package ‘foo.bar’                                                                                              
* removing ‘/usr/lib64/R/library/foo.bar’                                                                                                    
sh: rm: command not found 

At this time RStudio was already installed. 目前已经安装了RStudio So I logged into RStudio in my browser at port 8787 and tried install.packages in RStudio and every package installed fine. 因此,我在浏览器的端口8787上登录了RStudio ,并尝试在RStudio安装install.packages ,并安装了每个软件包。

I guess the take-home message is to install packages inside RStudio instead of inside R in the shell. 我想带回家的消息是要在RStudio内安装软件包,而不是在Shell中的R内安装软件包。

BTW I also cannot log in RStudio as root; 顺便说一句,我也不能以root用户身份登录RStudio。 says invalid username and password. 说无效的用户名和密码。

声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.

 
粤ICP备18138465号  © 2020-2024 STACKOOM.COM