bioconductor_docker
bioconductor_docker copied to clipboard
Any call to `ggplot2` in the current `bioconductor_docker:devel` image makes `R`/`Rstuido` **CRASH**
Describe the bug
With the current bioconductor_docker:devel
image a call to ggplot2
out of the box crashes R
/RStudio
with an error like this:
29 Apr 2021 12:10:12 [rsession-rstudio] ERROR session hadabend; LOGGED FROM: rstudio::core::Error {anonymous}::rInit(const rstudio::r::session::RInitInfo&) src/cpp/session/SessionMain.cpp:680
Upon plotting, the R
console freezes and crashes with above error upon entering Return
.
To Reproduce Steps to reproduce the behavior:
-
docker pull bioconductor/bioconductor_docker:devel
-
Call
docker run -d \ -p 127.0.0.1:8787:8787 \ -v /tmp:/tmp \ -e ROOT=TRUE \ -e DISABLE_AUTH=TRUE \ bioconductor/bioconductor_docker:devel
-
Use a browser and navigate to
localhost:8787
. -
In the
R
cosole installggplot2
withBiocManager::install("ggplot2", update = FALSE)
- UPDATING OTHERR
PACKAGES DOES NOT HELP -
Run the following (derived from
?ggplot2::ggplot
) - DO NOT SOURCE, but step through script (Shift+ Return):library(ggplot2) df <- data.frame(gp = factor(rep(letters[1:3], each = 10)), y = rnorm(30)) ds <- do.call(rbind, lapply(split(df, df$gp), function(d) { data.frame(mean = mean(d$y), sd = sd(d$y), gp = d$gp) })) ggplot(df, aes(gp, y)) + geom_point() + geom_point(data = ds, aes(y = mean), colour = 'red', size = 3)
Expected behavior Plot to appear
Desktop (please complete the following information):
- OS: Manjaro testing
- Browser chromium
- Version Chromium 90.0.4430.93 Arch Linux
I've investigated this a little. Just recording what i've found so far,
When I start up the docker image using the docker-compose
file, with docker-compose up
, it starts up as non-root rstudio
user.


Then, I tried your command to run as root while disabling auth. I installed ggplot2 using a binary repository I have (and not using the traditional, BiocManager::install() and compile step). This worked fine, and produced a result (as seen below). I've also attached my terminal to show how I started it up

But when I try your command exactly, I do get the error when I do the traditional BiocManager::install()
. I do get the error you are talking about,

So the problem here could be one of two things,
- The ggplot2 version that is being installed
and/or
- Why doesn't it produce the same error when I launch the image the standard way as non-root (shown in screenshot 1) ?
I don't have the answer right now but figured my investigation to add to more debugging.
Thank you for looking into this ...
The problem persists and has now become even more pressing: as the bioconductor
version has rolled over, the recent bioconductor_docker:latest
(not just bioconductor_docker:devel
) produces the issue!
Renders the otherwise great images practically useless for ggplot2
-involving workflows ...
BUT: the CURRENT (as 0f 2021/06/02) bioconductor_docker:devel
does NOT display the behavior in my hands, so it's now a remedy to move to that.
I'm pretty sure this problem is tied to the version of RStudio that is installed in the release image. Updating the RStudio version should fix the issue:
https://github.com/rstudio/rstudio/issues/9373#issuecomment-850647249
I've had the preview release installed on the bioconductor_docker:latest
image and that has been working for me, but it looks like RStudio just bumped up their "release" download now to include the fix:
https://www.rstudio.com/products/rstudio/release-notes/