xpas::load() hangs forever on some databases

[blinard-BIOINFO]

I observed this weird behaviour a few times when loading tens of thousands of DB for clapas.

0. xpas v0.3.1
1. I call xpas::load() on some rps database
2. it hangs forever, and never finishes to load.
3. this happens only on the the cluster AND NOT on my local machine (same database file, it loads in 2 seconds).

To reproduce on lamarck :

# load env

conda activate /ngs/linard/conda/
source /ngs/nromashchenko/load_env.sh

# build problematic database

/beegfs/linard/CLAPPAS/clapas_core/lib/xpas/build/xpas-build-aa-pos \
 --merge-branches \
-k 5 -o 5.0 \
--ar-binary "$(which raxml-ng)" \
--refalign /beegfs/linard/CLAPPAS/33208/pruned/3B9RA/Ax/A0_nx0_la.align \
--reftree /beegfs/linard/CLAPPAS/33208/pruned/3B9RA/Tx/T0_nx0_la.tree.raxml.bestTree.rerooted \
--ar-dir /beegfs/linard/CLAPPAS/33208/pruned/3B9RA/Dx/A0_nx0_la/k5_o5.00/AR \
--workdir /beegfs/linard/CLAPPAS/33208/pruned/3B9RA/Dx/A0_nx0_la/k5_o5.00 \
--uncompressed

# compile target xpas_example_stats, then :

xpas_example_stats /beegfs/linard/CLAPPAS/33208/pruned/3B9RA/Dx/A0_nx0_la/k5_o5.00/DB_k5_o5.0.rps

# will hang forever with 100% CPU usage and no RAM increase ... 

[blinard-BIOINFO]

@nromashchenko

TEST 1 :
0) modified CMakelist to link xpas_aa_pos in recipe xpas_example_stats

1. binary xpas_example_stats built on lamarck
2. copy to local desktop computer
3. ldd shows

	linux-vdso.so.1 (0x00007ffd269dd000)
	libboost_serialization.so.1.74.0 => not found
	libboost_iostreams.so.1.74.0 => not found
	libboost_system.so.1.74.0 => not found
	libboost_filesystem.so.1.74.0 => not found
	libz.so.1 => /lib/x86_64-linux-gnu/libz.so.1 (0x00007fa88bbc7000)
	libstdc++.so.6 => /lib/x86_64-linux-gnu/libstdc++.so.6 (0x00007fa88b9e5000)
	libm.so.6 => /lib/x86_64-linux-gnu/libm.so.6 (0x00007fa88b894000)
	libgcc_s.so.1 => /lib/x86_64-linux-gnu/libgcc_s.so.1 (0x00007fa88b879000)
	libc.so.6 => /lib/x86_64-linux-gnu/libc.so.6 (0x00007fa88b687000)
	/lib64/ld-linux-x86-64.so.2 (0x00007fa88bc01000)

4. installation of boost 1.74
5. execution on same database file => hangs forever on local desktop computer, behaviour of lamarck reproduced

CONCLUSION ==> it is the binary built on lamarck side that is faulty. Issue related to boost 1.74 ?

TEST 2 :
0) modified CMakelist to link xpas_aa_pos in recipe xpas_example_stats

1. xpas_example_stats build on local machine with local boost version (1.71).
2. ldd shows

	linux-vdso.so.1 (0x00007ffc5e9b9000)
	libboost_serialization.so.1.71.0 => /lib/x86_64-linux-gnu/libboost_serialization.so.1.71.0 (0x00007fa210dd7000)
	libboost_iostreams.so.1.71.0 => /lib/x86_64-linux-gnu/libboost_iostreams.so.1.71.0 (0x00007fa210db0000)
	libboost_filesystem.so.1.71.0 => /lib/x86_64-linux-gnu/libboost_filesystem.so.1.71.0 (0x00007fa210d92000)
	libstdc++.so.6 => /lib/x86_64-linux-gnu/libstdc++.so.6 (0x00007fa210bb0000)
	libgcc_s.so.1 => /lib/x86_64-linux-gnu/libgcc_s.so.1 (0x00007fa210b95000)
	libc.so.6 => /lib/x86_64-linux-gnu/libc.so.6 (0x00007fa2109a1000)
	libpthread.so.0 => /lib/x86_64-linux-gnu/libpthread.so.0 (0x00007fa21097e000)
	libz.so.1 => /lib/x86_64-linux-gnu/libz.so.1 (0x00007fa210962000)
	libbz2.so.1.0 => /lib/x86_64-linux-gnu/libbz2.so.1.0 (0x00007fa21094f000)
	liblzma.so.5 => /lib/x86_64-linux-gnu/liblzma.so.5 (0x00007fa210926000)
	libzstd.so.1 => /lib/x86_64-linux-gnu/libzstd.so.1 (0x00007fa21087d000)
	libm.so.6 => /lib/x86_64-linux-gnu/libm.so.6 (0x00007fa21072c000)
	/lib64/ld-linux-x86-64.so.2 (0x00007fa210e63000)

3. now it loads the database

• All linked libraries are the same, exepted : libpthread, liblzma, libzstd & libbz2
• why not on the server? . But that should not be an issue for this simple binary.