g++: internal compiler error: Killed (program cc1plus) - when build from Dockerfile

[32r81b]

I ran comand:

docker build https://raw.githubusercontent.com/kaldi-asr/kaldi/master/docker/ubuntu16.04-gpu/Dockerfile

In the end I got this output:

make[1]: Entering directory '/opt/kaldi/src/chain'
g++ -std=c++11 -I.. -isystem /opt/kaldi/tools/openfst-1.6.7/include -O1  -Wall -Wno-sign-compare -Wno-unused-local-typedefs -Wno-deprecated-declarations -Winit-self -DKALDI_DOUBLEPRECISION=0 -DHAVE_EXECINFO_H=1 -DHAVE_CXXABI_H -DHAVE_MKL -I/opt/intel/mkl/include -m64 -msse -msse2 -pthread -g -fPIC -DHAVE_CUDA -I/usr/local/cuda/include -fPIC -pthread -isystem /opt/kaldi/tools/openfst-1.6.7/include   -c -o chain-supervision.o chain-supervision.cc
g++: internal compiler error: Killed (program cc1plus)
Please submit a full bug report,
with preprocessed source if appropriate.
See <file:///usr/share/doc/gcc-5/README.Bugs> for instructions.
make[1]: *** [chain-supervision.o] Error 4
<builtin>: recipe for target 'chain-supervision.o' failed
make[1]: Leaving directory '/opt/kaldi/src/chain'
Makefile:136: recipe for target 'chain' failed
make: *** [chain] Error 2
The command '/bin/sh -c git clone --depth 1 https://github.com/kaldi-asr/kaldi.git /opt/kaldi &&     cd /opt/kaldi &&     cd /opt/kaldi/tools &&     ./extras/install_mkl.sh &&     make -j $(nproc) &&     cd /opt/kaldi/src &&     ./configure --shared --use-cuda &&     make depend -j $(nproc) &&     make -j $(nproc)' returned a non-zero code: 2

I appreciate any help.

[danpovey]

These things always due to too little memory. (OOM killer). Not sure if this is something that the dockerfile should be fixed for because people have different types of machine. cc @mdoulaty, but closing the issue as probably we wouldn't regard it as a bug. Look for the -j option given to Make, and reduce the numeric argument or remove the option.

[32r81b]

I have 16GB RAM on my local machine. In any case when I compale Kaldi on another machine all goes fine.

Thank you, Dan.

[32r81b]

It happend becouse i start Odcker container on Windows host machine. Releated dicussion docker/for-win#403

[mdoulaty]

How many CPU cores are visible to the Docker engine?
The image uses all of the available cores (that are available to the Docker engine) and it could well be that it’s running out of memory
You can try to explicitly set the number of cores for the make, replace ‘-j’ with ‘-j 1’ to see if it helps

[32r81b]

@mdoulaty , I fix this problem with recreating docker machine with biger memory. By default Docker machine start with a small amount of memory.

docker-machine rm default
docker-machine create -d virtualbox --virtualbox-cpu-count=6 --virtualbox-memory=12096 --virtualbox-disk-size=100000  default 

[Lancelof2019]

This issue happens when there is not enough memory.

https://bugzilla.redhat.com/show_bug.cgi?id=1199418