I am running some algorithm on my machine that requires a version of glibc 2.14 or above. For computing time improvements, I am migrating my whole code to the clusters of my university. However, the glibc version is 2.12 and they don't want to upgrade it. I am therefore trying to directly install the library on my allocated space and somehow link my algorithm to this library. I downloaded the 2.19 version.
I ran the following lines on my allocated space in the clusters
glibc-2.19/configure --prefix=/
make
make install
and I am facing the following issues when I am running the last command
/usr/bin/install -c -m 644 /home/myusername/libc.a //lib/libc.a
/usr/bin/install: cannot create regular file `//lib/libc.a': Permission denied
it's trying to create a file in a folder where I don't have permission but I don't know how to override that.
Any help is greatly appreciated!
glibc-2.19/configure --prefix=/
This is just wrong. You likely don't understand what --prefix does.
The default prefix is /usr/local. For system GLIBC, the prefix should most often be /usr.
What you likely want is --prefix=/home/$USER/glibc-install or something like that.
You should also read this answer.
Related
Since I installed according to the guide here on wsl2 ubuntu 20.04, I've been having errors related to libstc++.so.6, specifically GLIBCXX_3.4.26 not found (required by ...) where ... refers to different files within /opt/OpenFOAM/ThirdParty-v2006/platforms/linux64/gcc-6.3.0/lib64/ ending in .so, .so.1, .so.6 and so on (for instance, when running paraFoam the error would appear with respect to about 20 such files). I am able to successfully visualize the cavity tutorial (in paraview installation on windows).
I could get the errors to go away by doing what the user laborg suggested on Jan 4 for a similar problem with julia (see here), specifically copy libstdc++.so.6 from /usr/lib/x86_64-linux-gnu to /opt/OpenFOAM/ThirdParty-v2006/platforms/linux64/gcc-6.3.0/lib64/.
The questions is whether this copy-paste solution is recommended; will it come back and haunt me later? Is the libstdc++.so.6 from system installation going to be an issue if used in the lib64 folder of openfoam?
An additional info concerning openfoam installation, foamInstallationTest shows *not installed* errors against flex, wmake, gcc, g++, icoFoam and *critical error* for gcc, g++, icoFoam; but I as given here, foamInstallationTest is not meant for installation from the tar file. Openfoam installation seems to be alright based on the running of the cavity tutorial.
ok, please don't do copy past operation to solve this problem. The error means that you haven't installed the pre request libraries in your ubuntu. It seems that you have missed the first step in the tutorial.
It is not recommended but it will not hurt as long as the GLIBC versions returned from this command
strings /opt/OpenFOAM/ThirdParty-v2006/platforms/linux64/gcc-6.3.0/lib64/libstdc++.so.6 | grep GLIBC
are a subset of the GLIBC versions from this command.
strings /usr/lib/x86_64-linux-gnu/libstdc++.so.6 | grep GLIBC
which was no doubt the case for your Ubuntu setup.
A less risky route would be to redirect the soft link /opt/OpenFOAM/ThirdParty-v2006/platforms/linux64/gcc-6.3.0/lib64/libstdc++.so.6 to point to your other libstdc++.so.6 (that way you retain both versions)
ln -sf /usr/lib/x86_64-linux-gnu/libstdc++.so.6 /opt/OpenFOAM/ThirdParty-v2006/platforms/linux64/gcc-6.3.0/lib64/libstdc++.so.6
Then, if you hit an issue, you can always reset the link back to its original target. Of course /usr/lib/x86_64-linux-gnu/libstdc++.so.6 is itself a soft link, but you can point to it all the same or you can point to its target.
I believe the issue you are hitting is a derivative of the one mentioned here https://www.cfd-online.com/Forums/main/229027-persistence-glibcxx_3-4-26-not-found.html, which would point towards the fact that it is not an installation error on your part but an issue related to the packaging of the OpenFoam binaries. I agree it would screw up the wsl2 setup owing to the way OpenFoam prepends everything to paths. Of course the safest route is to compile from source using the Ubuntu system's gcc and thereby bypass the ThirdParty.
Seeing as you are using Ubuntu in the WSL instance, could also just install the Ubuntu package directly:
https://develop.openfoam.com/Development/openfoam/-/wikis/precompiled/debian
This problem comes from this line in the tutorial:
echo "source /opt/OpenFOAM/OpenFOAM-v2012/etc/bashrc" >> ~/.bashrc
This will point to OpenFOAM's libstdc++ everytime you open a terminal (or start a WSL2 session). If your workflow is not related to OpenFOAM, that can be an issue. If you remove or comment that line in your ~/.bashrc things should get back to normal. You can use nano in WSL2.
nano ~/.bashrc
Then comment:
#source /opt/OpenFOAM/OpenFOAM-v2012/etc/bashrc
However, as OpenFOAM uses that bashrc, you will need to source the OpenFOAM bashrc in each terminal before using openFOAM.
source /opt/OpenFOAM/OpenFOAM-v2012/etc/bashrc
My personal choice is to keep that line commented and, if I have a long work session using OpenFOAM, I just use nano to uncomment it, so every shell that I open works without sourcing again.
There are more elegant or complex approaches, but I prefer this one.
This answer should be valid with the 2006 version too, the link you shared points to 2012, so I guess they just updated the tutorial. If you installed 2006, just make sure when you source comment/uncomment to use the correct name.
In the same manner, if you followed another tutorial with another tool and sourced another library, you may experience issues.
Just start by taking a look at your bashrc and cleaning it.
I am trying to setup watchman for the mac. As stated on the website, I need to install glibtool.
Can anyone provide a link to where I can download glibtool?
I need to be able to download it from its source and the only solution I can find is by using brew.
I do not want to use brew.
Thank you.
glibtool is "GNU Libtool". It is typically installed as libtool on most systems, but because macOS has its own libtool that has completely different functionality, it is usually installed as glibtool on macOS.
If you can't directly use homebrew to install it, you can duplicate the steps in its recipe, which you can find here: https://github.com/Homebrew/homebrew-core/blob/master/Formula/libtool.rb
For the sake of keeping this answer "working" even if homebrew goes away, the homepage for libtool is https://www.gnu.org/software/libtool/ and you can follow the instructions there for information on how to build and install it.
I've covered similar issues to Wez's answer from a MacPorts perspective; I'll go ahead and assume you can't use that either.
The latest stable version at this time is 2.4.6. Typical best practice is to make a directory, e.g., build in the top level of the source. Add the prefix: g, with --program-prefix=g, the top level installation directory --prefix=PREFIX, or specify more fine-grained installation directories options for bin, include, lib, and share directories.
> mkdir build
> cd build
> ../configure --prefix=/my/install/path --program-prefix=g
> make; make install
You now have glibtool and glibtoolize in $PREFIX/bin.
I'm trying to help build a Ruby wrapper around Tensorflow using Swig. Currently, I'm stuck at making a shared build, .so, and exposing its C/C++ headers to Ruby. So the question is: How do I build a libtensorflow.so shared build including the full Tensorflow library so it's available as a shared library on OSX El Capitan (note: /usr/lib/ is read-only on El Capitan)?
Background
In this ruby-tensorflow project, I need to package a Tensorflow .bundle file, but whenever I irb -Ilib -rtensorflow or try to run the specs rspec, I get and errors that the basic numeric types are not defined, but they are clearly defined here.
I'm guessing this happens because my .so-file was not created properly or something is not linked as it should. C++/Swig/Bazel are not my strong sides, I'd like to focus on learning Tensorflow and building a good wrapper in Ruby, but I'm pretty stuck at this point getting to that fun part!
What I've done:
git clone --recurse-submodules https://github.com/tensorflow/tensorflow
cd tensorflow
bazel build //tensorflow:libtensorflow.so (wait 10-15min on my machine)
Copied the generated libtensorflow.so (166.6 MB) to the /ext-folder
Run the ruby extconf.rb, make, and make install described in the project
Run rspec
In desperation, I've also gone through the official installation from source several times, but I don't know if that, the last sudo pip install /tmp/tensorflow_pkg/tensorflow-0.9.0-py2-none-any.whl-step even creates a shared build or just exposes a Python interface.
The guy, Arafat, who made the original repository and made the instructions that I've followed, says his libtensorflow.so is 4.5 GB on his Linux machine – so over 20X the size of the shared build on my OSX machine. UPDATE1: he says his libtensorflow.so-build is 302.2 MB, 4.5GB was the size of the entire tensorflow folder.
Any help or alternative approaches are very appreciated!
After more digging around, discovering otool (thanks Kristina) and better understanding what a .so-file is, the solution didn't require much change in my setup:
Shared Build
# Clone source files
git clone --recurse-submodules https://github.com/tensorflow/tensorflow
cd tensorflow
# Build library
bazel build //tensorflow:libtensorflow.so
# Copy the newly shared build/library to /usr/local/lib
sudo cp bazel-bin/tensorflow/libtensorflow.so /usr/local/lib
Calling from Ruby using Swig
Follow the steps here, https://github.com/chrhansen/ruby-tensorflow#install-ruby-tensorflow, to run Swig, create a Makefile and make
When you run make you should see a line saying:
$ make
$ linking shared-object libtensorflow.bundle
If your shared build is not accessible you'll see something like:
$ ld: library not found for -ltensorflow
Simple tutorial
For those starting on this adventure, using C/C++ libraries in Ruby, this post was a good tutorial for me: http://engineering.gusto.com/simple-ruby-c-extensions-with-swig/
I don't think you actually want a .so, I think you want a .dylib (see What are the differences between .so and .dylib on osx?). You're forcing Bazel to build a .so by specifying libtensorflow.so as the target, build this instead:
bazel build //tensorflow
(//tensorflow is shorthand for //tensorflow:tensorflow, which is "build the tensorflow target." Specifying an exact file you want forces Bazel to build that file, if possible.)
Once you have a .dylib, you can check its contents with otool:
otool -L bazel-bin/tensorflow/libtensorflow.dylib
Not sure if this will solve all your problems, but worth a try.
I'm trying to install Armadillo (and thus also BLAS and LAPACK) on a linux server for which I do not have root permission. I have jumped through a few of the first hurdles, but I am getting an error:
/usr/bin/ld: cannot find -llapack
This question addresses the same problem, with the solution being to install
liblapack-dev
liblapack3
libopenblas-base
libopenblas-dev
The problem is the only way I've found to do this online is by doing something like sudo-aptget install or yum install. Both are not allowed on the server I use. I can download binaries and install them locally - that is it.
My question is: how can I install these packages without the above permissions and get on my way to using Armadillo?
I'm using Centos linux, if it helps.
Since you do not have root permissions, the best way is to download and build LAPACK and BLAS. You can download the source code from netlib.
Description of installation instructions can be found here and here.
The basic steps are:
Unzip and tar the file.
Copy and edit the file LAPACK/make.inc.example to LAPACK/make.inc.
Edit the file LAPACK/Makefile
type make.
Similarly you can download and build BLAS.
Is there a way to get a list of filenames/paths that make install copies to the filesystem? Some packages come with a MANIFEST file, but not the ones that I am working with.
I was just investigating this myself while compiling a custom version of QEMU. I used the following method to work out what was installed and where (as well as using it as a basis for a .deb file):
mkdir /tmp/installer
./configure --target-list=i386-softmmu
make
sudo make install DESTDIR=/tmp/installer
cd /tmp/installer
tree .
Tree is a utility that recursively displays the contents of a directory in a visually appealing manner - sudo apt-get install tree for Debian / Ubuntu users
Hope that helps someone... it took me a bit of poking around to nut it out, but I found it quite a useful way of visualising what was going on.
The most fool-proof way is to use chroot: have "make install" run inside a chroot jail; compute a list of the files that you had before the installation, and compare that to the list of files after the installation.
Many installations will support either a --prefix configuration option, and/or a DESTDIR environment variable. You can use those for a lighter-wait version of chroot (trusting that the installation will fail if it tries to write to a location outside these if you run installation as a fairly unprivileged user).
Another approach is to replace the install program. Many packages support an INSTALL environment variable that, well, is the install program to use; there are tracing versions of install around.
make uninstall might show the files as it removes them if the author of the compiling instructions provides the information to allow an uninstall (it has been awhile since I have done one so I can't say for sure).
Also make -n install will do a "dry run" of the install process and it may be reasonable to extract the information from its results.
It differs for every project that you run 'make install' on. The files which are installed are controlled by the install target in the Makefile being used. Your best bet is to open the Makefile and search for 'install:' - from there you can see what files will be copied out to your system.
Take a snapshot of the contents of the install location before installing
Install
Compare the current contents with the old contents.
Example:
./configure --prefix /usr/local
make -j`nproc`
find /usr/local | sort -u > /tmp/snapshot1
make install
find /usr/local | sort -u > /tmp/snapshot2
comm -3 /tmp/snapshot{1,2} # this prints the files added by `make install` to stdout
If the install program you're using doesn't support DESTDIR or --prefix (or an equivalent), I have found that it may be possible to identify new files as follows:
Start with as clean a system as possible (a fresh VM image is preferable)
Compile the software, wait a few minutes.
Install the software package.
Find files modified within the past 5 minutes: sudo find / -mmin -5 -type f (the find command has a ton of parameters for querying based on file modification / creation times, but this worked pretty well for me; you just need to narrow the time span so that you pick up the files created by the installer but nothing else).