unable to load shared object '/opt/conda/lib/R/library/sf/libs/sf.so

[btupper]

Hi,

Hmmm. Perhaps a path issue?

library(xyzt)
library(sf)

x <- xyzt::read_calcofi()

x <- sf::st_as_sf(x, coords = c("lon", "lat"),
                  dim = 'XY', crs = 4326)

#Error: package or namespace load failed for ‘sf’ in dyn.load(file, DLLpath = DLLpath, ...):
# unable to load shared object '/opt/conda/lib/R/library/sf/libs/sf.so':
#  /usr/lib/x86_64-linux-gnu/libstdc++.so.6: version `GLIBCXX_3.4.30' not found (required by /opt/conda/lib/R/library/sf/libs/../../../../libgdal.so.31)

# > traceback()
# 10: stop(msg, call. = FALSE, domain = NA)
# 9: value[[3L]](cond)
# 8: tryCatchOne(expr, names, parentenv, handlers[[1L]])
# 7: tryCatchList(expr, classes, parentenv, handlers)
# 6: tryCatch({
#   attr(package, "LibPath") <- which.lib.loc
#   ns <- loadNamespace(package, lib.loc)
#   env <- attachNamespace(ns, pos = pos, deps, exclude, include.only)
# }, error = function(e) {
#   P <- if (!is.null(cc <- conditionCall(e))) 
#     paste(" in", deparse(cc)[1L])
#   else ""
#   msg <- gettextf("package or namespace load failed for %s%s:\n %s", 
#                   sQuote(package), P, conditionMessage(e))
#   if (logical.return && !quietly) 
#     message(paste("Error:", msg), domain = NA)
#   else stop(msg, call. = FALSE, domain = NA)
# })
# 5: library(sf) at .active-rstudio-document#2
# 4: eval(ei, envir)
# 3: eval(ei, envir)
# 2: withVisible(eval(ei, envir))
# 1: source("~/.active-rstudio-document")
# > sessionInfo()
# R version 4.1.3 (2022-03-10)
# Platform: x86_64-conda-linux-gnu (64-bit)
# Running under: Debian GNU/Linux 11 (bullseye)
# 
# Matrix products: default
# BLAS/LAPACK: /opt/conda/lib/libopenblasp-r0.3.21.so
# 
# locale:
#   [1] LC_CTYPE=C.UTF-8       LC_NUMERIC=C           LC_TIME=C.UTF-8        LC_COLLATE=C.UTF-8    
# [5] LC_MONETARY=C.UTF-8    LC_MESSAGES=C.UTF-8    LC_PAPER=C.UTF-8       LC_NAME=C             
# [9] LC_ADDRESS=C           LC_TELEPHONE=C         LC_MEASUREMENT=C.UTF-8 LC_IDENTIFICATION=C   
# 
# attached base packages:
#   [1] stats     graphics  grDevices utils     datasets  methods   base     
# 
# other attached packages:
#   [1] xyzt_0.1
# 
# loaded via a namespace (and not attached):
#   [1] Rcpp_1.0.9         rstudioapi_0.13    magrittr_2.0.3     units_0.8-0        tidyselect_1.1.2  
# [6] R6_2.5.1           rlang_1.0.4        fansi_1.0.3        dplyr_1.0.9        tools_4.1.3       
# [11] ghrsst_0.1         grid_4.1.3         KernSmooth_2.23-20 utf8_1.2.2         cli_3.3.0         
# [16] e1071_1.7-11       DBI_1.1.3          class_7.3-20       assertthat_0.2.1   tibble_3.1.8      
# [21] lifecycle_1.0.1    purrr_0.3.4        vctrs_0.4.1        ncdf4_1.19         glue_1.6.2        
# [26] proxy_0.4-27       compiler_4.1.3     pillar_1.8.0       generics_0.1.3     classInt_0.4-7    
# [31] pkgconfig_2.0.3 
  

[abkfenris]

@ocefpaf Is it me or does it look like it's yelling about missing a specific version of libc?

[btupper]

I just noticed a warning that appears when the markdown is rendered, but doesn't appear when the same commands (above) are run at the R prompt. I have pasted it in below. It points to proj as an issue.

The markdown actually does render successfully, but coordinate reference system are absent (which is possible in sf).

x <- xyzt::read_calcofi()

x <- sf::st_as_sf(x, coords = c("lon", "lat"),
                  dim = 'XY', crs = 4326)

## Warning in CPL_crs_from_input(x): GDAL Error 1: PROJ: proj_create_from_database:
## Open of /opt/conda/share/proj failed

[ocefpaf]

@ocefpaf Is it me or does it look like it's yelling about missing a specific version of libc?

@abkfenris looks like our Python image is OK b/c we can not run an older glibc on a newer they are all forward compatible. (The opposite is not true though.)

Could this be related to RStudio? After I chose the Python image on the hub I'm having issue to switch to the RStudio one, so I cannot test this there at the moment.

[abkfenris]

https://oceanhackweek.2i2c.cloud/hub/home to stop your server and switch to the R profile

[abkfenris]

In the R profile

(base) jovyan@jupyter-abkfenris:~$ /sbin/ldconfig -p | grep stdc++
        libstdc++.so.6 (libc6,x86-64) => /usr/lib/x86_64-linux-gnu/libstdc++.so.6
        libstdc++.so.6 (libc6) => /usr/lib32/libstdc++.so.6
(base) jovyan@jupyter-abkfenris:~$ strings /usr/lib/x86_64-linux-gnu/libstdc++.so.6 | grep LIBCXX
bash: strings: command not found

[ocefpaf]

Yep. Just did that and....

 /sbin/ldconfig -p | grep stdc++        libstdc++.so.6 (libc6,x86-64) => /usr/lib/x86_64-linux-gnu/libstdc++.so.6
        libstdc++.so.6 (libc6) => /usr/lib32/libstdc++.so.6

I'm not sure there are two. The images are not based on the same base are they? The strings command is not available to explore more. Looking into it...

[ocefpaf]

@btupper I'm not familiar with R packaging but could any of the dependencies we are installing bring a pre-compiled binary? Could that be the culprit?

---

Probably not. All the GitHub installs are using dependencies=FALSE, upgrade_dependencies=FALSE, upgrade=FALSE.

[ocefpaf]

Hmmm. Perhaps a path issue?

Most likely. I just tried the RStudio console and it breaks as you reported:

But if I open a terminal, then R, then try that code, everything works:

Not sure how to change the lookup path for RStudio, or how it would break if we do so!

[btupper]

RStudio -> markdown rendering: throws error as shown above
RStudio -> console: throws warning abut proj, not an error, CRS lost
RStudio -> Terminal: now error, no warning, CRS preserved - all good!

R console (via jupyter-hub/lab): throws proj warning, complains of missing geojsonio package

[ocefpaf]

oceanhackweek/jupyter-image#57 kind of works for me. The proj data still requires user to do Sys.setenv(PROJ_LIB="/opt/conda/share/proj") before running but at least the code above works as expected.

[abkfenris]

RStudio seems to be coming up it's own environment.

> Sys.getenv()
CLICOLOR_FORCE             1
DISPLAY                    :0
EDITOR                     vi
GIT_ASKPASS                rpostback-askpass
HOME                       /home/jovyan
LANG                       C.UTF-8
LD_LIBRARY_PATH            /opt/conda/lib/R/lib:/opt/conda/lib:/opt/conda/lib
LN_S                       ln -s
LOGNAME                    jovyan
MAKE                       make
MPLENGINE                  tkAgg
PAGER                      less
PATH                       /usr/lib/rstudio-server/bin/:/opt/conda/lib/R/library/littler/bin:/opt/conda/bin:/opt/conda/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/lib/rstudio-server/bin/quarto/bin:/usr/lib/rstudio-server/bin/postback:/usr/lib/rstudio-server/bin/postback
R_BROWSER                  
R_BZIPCMD                  /opt/conda/bin/bzip2
R_CLI_HAS_HYPERLINK_IDE_HELP
                           true
R_CLI_HAS_HYPERLINK_IDE_RUN
                           true
R_CLI_HAS_HYPERLINK_IDE_VIGNETTE
                           true
R_DOC_DIR                  /opt/conda/lib/R/doc
R_GZIPCMD                  gzip
R_HOME                     /opt/conda/lib/R
R_INCLUDE_DIR              /opt/conda/lib/R/include
R_LIBS_SITE                
R_LIBS_USER                ~/R/x86_64-conda-linux-gnu-library/4.1
R_PAPERSIZE                a4
R_PDFVIEWER                
R_PLATFORM                 x86_64-conda-linux-gnu
R_PRINTCMD                 
R_RD4PDF                   times,inconsolata,hyper
R_SESSION_TMPDIR           /tmp/RtmpAaZMU4
R_SHARE_DIR                /opt/conda/lib/R/share
R_STRIP_SHARED_LIB         x86_64-conda-linux-gnu-strip --strip-unneeded
R_STRIP_STATIC_LIB         x86_64-conda-linux-gnu-strip --strip-debug
R_SYSTEM_ABI               linux,gcc,gxx,gfortran,gfortran
R_TEXI2DVICMD              /home/conda/feedstock_root/build_artifacts/r-base-split_1660164760553/_build_env/bin/texi2dvi
R_UNZIPCMD                 
R_ZIPCMD                   
RMARKDOWN_MATHJAX_PATH     /usr/lib/rstudio-server/resources/mathjax-27
RS_RPOSTBACK_PATH          /usr/lib/rstudio-server/bin/rpostback
RS_SERVER_RPC_SOCKET_PATH
                           /var/run/rstudio-server/rstudio-rserver/session-server-rpc.socket
RS_SESSION_TMP_DIR         /var/run/rstudio-server/rstudio-rsession
RSTUDIO                    1
RSTUDIO_CLI_HYPERLINKS     true
RSTUDIO_CONSOLE_COLOR      256
RSTUDIO_CONSOLE_WIDTH      93
RSTUDIO_HTTP_REFERER       http://127.0.0.1:8080/rstudio/
RSTUDIO_PANDOC             /usr/lib/rstudio-server/bin/quarto/bin/tools
RSTUDIO_PROGRAM_MODE       server
RSTUDIO_R_MODULE           
RSTUDIO_R_PRELAUNCH_SCRIPT
                           
RSTUDIO_R_REPO             
RSTUDIO_R_VERSION_LABEL    
RSTUDIO_SESSION_PID        75
RSTUDIO_SESSION_STREAM     jovyan-d
RSTUDIO_USER_IDENTITY      jovyan
RSTUDIO_USER_IDENTITY_DISPLAY
                           jovyan
RSTUDIO_WINUTILS           bin/winutils
SED                        /opt/conda/bin/sed
SSH_ASKPASS                rpostback-askpass
TAR                        tar
TERM                       xterm-256color
USER                       jovyan

(base) $ env
CONDA_PROMPT_MODIFIER=(base) 
CONDA_ENV=base
JUPYTER_SERVER_ROOT=/home/jovyan
HOSTNAME=b0207e79e72e
LD_LIBRARY_PATH=/opt/conda/lib
HOME=/home/jovyan
CONDA_SHLVL=1
UDUNITS2_XML_PATH=/opt/conda/share/udunits/udunits2.xml
PYTHONUNBUFFERED=1
CPL_ZIP_ENCODING=UTF-8
NB_USER=jovyan
_CE_M=
RSTUDIO_WHICH_R=/opt/conda/bin/R
PYXTERM_DIMENSIONS=80x25
PYTHONDONTWRITEBYTECODE=1
R_LIBS_USER=/opt/r
TERM=xterm-256color
COLUMNS=80
_CE_CONDA=
RSTUDIO_URL=https://download2.rstudio.org/server/bionic/amd64/rstudio-server-2022.07.1-554-amd64.deb
_CONDA_SET_PROJ_LIB=/opt/conda/share/proj
PATH=/opt/conda/bin:/opt/conda/condabin:/usr/local/bin:/usr/bin:/bin:/usr/local/games:/usr/games
PROJ_NETWORK=ON
LANG=C.UTF-8
CONDA_PYTHON_EXE=/opt/conda/bin/python
GSETTINGS_SCHEMA_DIR=/opt/conda/share/glib-2.0/schemas
GDAL_DRIVER_PATH=/opt/conda/lib/gdalplugins
CONDA_DEFAULT_ENV=base
CONDA_DIR=/opt/conda
LC_ALL=C.UTF-8
PROJ_LIB=/opt/conda/share/proj
PWD=/home/jovyan
CONDA_EXE=/opt/conda/bin/conda
PYDEVD_USE_FRAME_EVAL=NO
LINES=25
NB_UID=1000
CONDA_PREFIX=/opt/conda
GSETTINGS_SCHEMA_DIR_CONDA_BACKUP=
GDAL_DATA=/opt/conda/share/gdal
JUPYTER_SERVER_URL=http://b0207e79e72e:8080/

[jevanilla]

@ocefpaf I'm still getting this error in the RStudio instance on the hub:

Sys.setenv(PROJ_LIB="/opt/conda/share/proj")
library(sf)

Error: package or namespace load failed for ‘sf’ in dyn.load(file, DLLpath = DLLpath, ...): unable to load shared object '/opt/conda/lib/R/library/sf/libs/sf.so': /usr/lib/x86_64-linux-gnu/libstdc++.so.6: version `GLIBCXX_3.4.30' not found (required by /opt/conda/lib/R/library/sf/libs/../../../../libgdal.so.31)

[abkfenris]

These changes haven't gone live yet. The image hasn't even finished building! https://github.com/oceanhackweek/jupyter-image/runs/7860496038?check_suite_focus=true

[abkfenris]

Ok, the image is built, but we need to wait for 2i2c to approve and merge the new image: 2i2c-org/infrastructure#1631

[abkfenris]

Now for 2i2c's deployment, then you can shut down and restart your profile to test https://github.com/2i2c-org/infrastructure/actions/runs/2869052365

[abkfenris]

Now it's yelling about being linked to the wrong GDAL GEOS linking again

[abkfenris]

Waiting on a new PR to be merged: 2i2c-org/infrastructure#1632 that will hopefully get us down to just warnings after adding a Rprofile.site in oceanhackweek/jupyter-image#59

[btupper]

Sweet - my thanks to you and @ocefpaf for lifting us over this bump!

[abkfenris]

We still get the warnings about GEOS, but at least proj is playing nice.