makedb (step=1) stuck for -p=diamond sometimes
approx. 1/10 runs will fail within step=2 for QfO_2020 (78 fasta files) using 64 cores:
proteinortho6.pl (v6.1.3):
my $makedb_ret = `$cmd 2>&1`;
code for testing:
rm *err;
for i in $(seq 1 10); do
rm -rf input_$i* 2>/dev/null;
cp -r input input_$i;
/usr/bin/time -f 'real time(s)\tmax resident set size(KB)\tCPUusage\tKerneltime\tUsertime\tMemoryusage\n%e\t%M\t%P\t%S\t%U\t%K' timeout 300 /home/paul/Documents/proteinortho_dev/REBOOT/bin/proteinortho_v6/proteinortho6.pl -debug=1 -step=1 -cpus=64 -project=input_$i input_$i/*fasta >$i.err 2>&1 &
done
observe output : tail -n3 *err
:
fish for unjoined threads in output 3:
grep "generate_" 3.err | grep spawn | grep -v -E $(grep "generate_" 3.err | grep dead | sed -E 's/.*(:[0-9]+).*/\1/g' | tr '\n' '|' | sed 's/|$//#')
-> then grep for the last fasta file, failes at the $cmd
call.
Edited by Paul Klemm