Segmentation fault (core dumped)

[gonzaparra]

Hi all!

I am trying to reinstall a package that I was previously able to install and use. I was building a package of my own after my computer unexpectedly restarted and then I started to have problems loading the rpgraph package. So I decided to uninstall it and to reinstall it. When I did so I got the following error:

`library(devtools)

library(rJava)
install_github("Albluca/rpgraph")

Downloading GitHub repo Albluca/rpgraph@master
from URL https://api.github.com/repos/Albluca/rpgraph/zipball/master
Installing rpgraph
Running command /usr/lib/R/bin/R
Arguments:
CMD
INSTALL
/tmp/Rtmp5OrtLL/devtools505a703b3ccd/Albluca-rpgraph-de04f96
--library=/home/gonzalo/R/x86_64-pc-linux-gnu-library/3.4
--install-tests

• installing source package ‘rpgraph’ ...
  ** R
  ** data
  *** moving datasets to lazyload DB
  ** inst
  ** preparing package for lazy loading
  ** help
  *** installing help indices
  ** building package indices
  ** installing vignettes
  ** testing if installed package can be loaded
  Segmentation fault (core dumped)
  ERROR: loading failed
• removing ‘/home/gonzalo/R/x86_64-pc-linux-gnu-library/3.4/rpgraph’
  Installation failed: run(bin, args = real_cmdargs, stdout_line_callback = real_callback(stdout), stderr_line_callback = real_callback(stderr), stdout_callback = real_block_callback, stderr_callback = real_block_callback, echo_cmd = echo, echo = show, spinner = spinner, error_on_status = fail_on_status, timeout = timeout) : System command error`

I tried reinstalling JAva and cran from zero but somehow there is something that remains in my system that avoids the package to be installed. Also, since the error is not very informative, I have no idea on how to narrow where the problem is.

Thanks for any help you can provide.

[colearendt]

Having the same issue in the xlsx package. Fairly confident it is related to security updates that were released for Ubuntu in the last few days (6/19, I think). I fortunately had not installed the updates on my computer yet when the issue came in, so I was able to confirm that installing the updates is what broke the package install.

Further, I have confirmed that the .jpackage function is what is causing the segfault. It happens whether or not I use .jpackage('xlsx') or .jpackage('xlsxjars'), which have different jars included (so I do not suspect the Java code in particular). To make matters more interesting, the segfault only seems to occur when .jpackage is called during .onLoad or .onAttach, and specifically when the package is being installed - calling it manually works after the package is installed.

I would suggest that this is a fairly high priority bug, as I would expect it to affect all Ubuntu OS's and all packages that use rJava with jpackage called in .onLoad or .onAttach.

[cs-dieter-kling]

I would suggest that this is a fairly high priority bug, as I would expect it to affect all Ubuntu OS's and all packages that use rJava with jpackage called in .onLoad or >.onAttach.

Absolutely! Segmentation fault bug affects also RJDBC - we cannot access our Hive DB any more after latest OS updates (incl. linux kernel).
It tested with latest Debian, Centos, Ubuntu OSes and JDK 8 (Oracle JDK and OpenJDK) the error persists.
A workaround seems to be using older JDK version 1.7.0_95 but I'm still testing...
This does not work either :-( or to be precise RDJBC still fails in 57% of cases with Segmentation fault

[colearendt]

Workaround used for the xlsx package was to temporarily suspend calling .jpackage from .onLoad / .onAttach when grepl('linux-gnu',R.version$os). This is less than ideal because it requires the user then to execute .jpackage('xlsx') manually - but hopefully not too much to ask if they are just now installing the package.

An avenue for exploring the root cause might be to pull apart the .jpackage function and execute line by line within .onLoad / .onAttach

[colearendt]

Very interesting. Looks like the seg fault comes from .jinit() within installing a package for some reason - I do not know enough about R / Java internals to tell why - in particular, this piece of the .jpackages function:

  if (!.jniInitialized) 
    .jinit()

Calling .jinit() from an interactive session has no issues.

[colearendt]

In particular, looks to be this line of .jinit():

xr <- .External(RinitJVM, boot.classpath, parameters)

Will need someone who is more versed than I in C/C++ to dig into RinitJVM to see what the issue is. Not sure how to nicely reproduce during install or if there is a way to diagnose by executing partial files / etc.. For testing, I have been building binary files locally (which will reproduce the segfault on Ubuntu) by running successive lines of code within .onAttach.

[gonzaparra]

mmm interesting but I have not updated any package in my OS so how can it be related to it?

But yeah something weird is going on, since I am not really able to solve it. It works on other computers though so I guess that if I cannot solve it I will just need to make an OS reinstallation without updates maybe to be safe.

[colearendt]

@gonzaparra i would recommend forking the repo and removing any line like .jinit or .jpackage from the .onAttach or .onLoad functions and see if that allows you to install. You should be able to call the functions yourself after installation and loading.

The restart you did probably allowed the problematic security updates to take effect

[gonzaparra]

Thanks for your answers. I already reinstalled my OS. I installed the latest Ubuntu LTS version and now everything works.. I hope these updated will be fixed cause is very problematic. Thanks for your time and answers if it happens again I hope I will be able to solve it without formatting my pc. Best Gonza El 22 jun. 2017 18:07, "Cole Arendt" <notifications@github.com> escribió:

…

@gonzaparra <https://github.com/gonzaparra> i would recommend forking the repo and removing any line like .jinit or .jpackage from the .onAttach or .onLoad functions and see if that allows you to install. You should be able to call the functions yourself after installation and loading. The restart you did probably allowed the problematic security updates to take effect — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#110 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AGDtRE0_ymI9f3t-uV4Jm_mmCJRodYh-ks5sGpEqgaJpZM4OBQaB> .

[colearendt]

@s-u Not sure if you have seen, but this is a crippling bug for rJava on UNIX platforms related to a kernel change related to heap memory in the past week. Best I can tell, the problem comes from

xr <- .External(RinitJVM, boot.classpath, parameters)

Update from the xlsx package: my workaround allows install / library / .jpackage or .jinit, but then close R session / open new R session / library / .jpackage or .jinit fails.

[cs-dieter-kling]

@colearendt @s-u Not sure if you have seen, but this is a crippling bug for rJava on UNIX >platforms related to a kernel change related to heap memory in the past week.

I can confirm, in our case RJDBC fails with Segmentation fault after security updtates of the Linux Kernel
Old version:
Linux 3.16.0-4-amd64 #1 SMP Debian 3.16.39-1 (2016-12-30) x86_64 GNU/Linux
new Kernel version:
Linux 3.16.0-4-amd64 #1 SMP Debian 3.16.43-2+deb8u1 (2017-06-18) x86_64 GNU/Linux

lsb_release -a:

No LSB modules are available.
Distributor ID: Debian
Description:    Debian GNU/Linux 8.8 (jessie)
Release:    8.8
Codename:   jessie

Link to Debian Security Advisory - https://www.debian.org/security/2017/dsa-3886

[cs-dieter-kling]

Increasing stack size of the JVM by setting -Xss2560k Java option did the trick for me
in order to get RJDBC back to work:

options( java.parameters = c("-Xss2560k", "-Xmx2g") )
...

library(rJava)
library(RJDBC)

...

.jinit(classpath=c(hive.jar.path, hadoop.jar.path))

...

There are couple of workarounds have been suggested in order to counteract Segmentation faults:

• the above mentioned workaround - https://www.kvraudio.com/forum/viewtopic.php?p=6813554
• disabling Address Space Layout Randomization - 0.17 - 0.19 segfaults on powerpc64le some of the time, can't reproduce in gdb rust-lang/cargo#4197 , https://askubuntu.com/questions/318315/how-can-i-temporarily-disable-aslr-address-space-layout-randomization

But it would be interesting to know why the bug fix of Stack Clash causes such issues and how to fix the root cause.

[colearendt]

No luck increasing stack size for the xlsx package, at least not for me - tried up to 10240k. Other workarounds were a no-go as well, unfortunately.

I definitely agree that root cause would be great to know, as such debilitating issues are unsettling.