You can generate a core file to find where the problem originates. In the backstrace below, none of the addresses is specific to Ray.
On 09/03/15 05:09 AM, Xiaofen Wu wrote: > Hello, > > I am recently using Ray meta for metagenome assembly. And I used > different kmer-size (from 31 to 81), it worked always fine, but this > time it crashed in halfway, and its not because run out the memory. The > error message I got is like this: > > [m119:26639] *** Process received signal *** > > [m119:26639] Signal: Segmentation fault (11) > > [m119:26639] Signal code: Address not mapped (1) > > [m119:26639] Failing at address: 0x88 > > [m119:26639] [ 0] /lib64/libpthread.so.0(+0xf710)[0x2abfc26ed710] > > [m119:26639] [ 1] > /opt/openmpi/1.7.5gcc4.4/lib/openmpi/mca_oob_tcp.so(+0xbd48)[0x2abfc4c37d48] > > [m119:26639] [ 2] > /opt/openmpi/1.7.5gcc4.4/lib/openmpi/mca_oob_tcp.so(mca_oob_tcp_peer_recv_connect_ack+0x87)[0x2abfc4c37fb7] > > [m119:26639] [ 3] > /opt/openmpi/1.7.5gcc4.4/lib/openmpi/mca_oob_tcp.so(+0x7c20)[0x2abfc4c33c20] > > [m119:26639] [ 4] > /opt/openmpi/1.7.5gcc4.4/lib/libopen-pal.so.6(opal_libevent2021_event_base_loop+0x8bc)[0x2abfc2f709dc] > > [m119:26639] [ 5] > /opt/openmpi/1.7.5gcc4.4/lib/libopen-rte.so.7(+0x38b1e)[0x2abfc2cc8b1e] > > [m119:26639] [ 6] /lib64/libpthread.so.0(+0x79d1)[0x2abfc26e59d1] > > [m119:26639] [ 7] /lib64/libc.so.6(clone+0x6d)[0x2abfc29e49dd] > > [m119:26639] *** End of error message *** > > -------------------------------------------------------------------------- > > mpiexec noticed that process rank 1 with PID 26639 on node m119 exited > on signal 11 (Segmentation fault). > > -------------------------------------------------------------------------- > > slurmstepd: _slurm_cgroup_destroy: problem deleting step cgroup path > /cgroup/freezer/slurm/uid_44133/job_4258126/step_batch: Device or > resource busy > > > The following command is what I used. Does anyone have any idea why it > crashed? > > for k in `seq 31 10 81`; do > > mpiexec -n 32 Ray -k $k -p qtrim1.fastq qtrim2.fastq -o out_$k > > > I also have another problem using the above command. When I was running > Ray, I got weird log file sometimes, like below. Those lines start with > Rank are not supposed to appear there, because normally the log file > should start from Ray LICENSE part. I checked everything, but this did > not happen to all samples or all kmer size, only happen to one or two > kmer size. I am really wondering why it happened, and I am a bit worried > about the results, I cannot be certain if it was misassembled. Do you > think its something to do with my command? Thanks in advance! > > > Rank 1: assembler memory usage: 34460 KiB > > Rank 2: assembler memory usage: 34464 KiB > > Rank 4: assembler memory usage: 34460 KiB > > Rank 5: assembler memory usage: 34464 KiB > > Rank 6: assembler memory usage: 34388 KiB > > Rank 7: assembler memory usage: 34388 KiB > > Rank 9: assembler memory usage: 34388 KiB > > Rank 10: assembler memory usage: 34388 KiB > > Rank 11: assembler memory usage: 34460 KiB > > Rank 12: assembler memory usage: 34388 KiB > > Rank 13: assembler memory usage: 34388 KiB > > Rank 15: assembler memory usage: 34460 KiB > > Rank 3: assembler memory usage: 34464 KiB > > > ************************************************** > > This program comes with ABSOLUTELY NO WARRANTY. > > This is free software, and you are welcome to redistribute it > > under certain conditions; see "LICENSE.txt" for details. > > ************************************************** > > > Ray Copyright (C) 2010, 2011, 2012, 2013 Sébastien Boisvert > > Centre de recherche en infectiologie de l'Université Laval > > Project funded by the Canadian Institutes of Health Research (Doctoral > award 200902CGM-2 > > > Best > > Xiaofen > > > > ------------------------------------------------------------------------------ > Dive into the World of Parallel Programming The Go Parallel Website, sponsored > by Intel and developed in partnership with Slashdot Media, is your hub for all > things parallel software development, from weekly thought leadership blogs to > news, videos, case studies, tutorials and more. Take a look and join the > conversation now. http://goparallel.sourceforge.net/ > > > > _______________________________________________ > Denovoassembler-users mailing list > Denovoassembler-users@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/denovoassembler-users > ------------------------------------------------------------------------------ Dive into the World of Parallel Programming The Go Parallel Website, sponsored by Intel and developed in partnership with Slashdot Media, is your hub for all things parallel software development, from weekly thought leadership blogs to news, videos, case studies, tutorials and more. Take a look and join the conversation now. http://goparallel.sourceforge.net/ _______________________________________________ Denovoassembler-users mailing list Denovoassembler-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/denovoassembler-users
