Re: parallel + blast + LSF

[Ole Tange]

On Wed, Apr 15, 2015 at 3:34 PM, Giuseppe Aprea
<giuseppe.aprea@gmail.com> wrote:

> I am trying to use GNU parallel v. 20150122 with blast for a very large
> sequences alignment. I am using Parallel on a cluster which uses LSF as
> queue system.

I have never run anything on a LSF system, so take my advice with 1
mmol of NaCl.

> "servers" is this file:
>
> /afs/enea.it/software/bin/blaunch.sh cresco3x013.portici.enea.it
> /afs/enea.it/software/bin/blaunch.sh cresco3x013.portici.enea.it
:

Duplicate lines in a --slf file are merged. It does no harm to have
the duplicate lines, but the duplicates are simply merged into 1.

> My problems are
:
> - the result files are empy and I can see the following messages:

It has been a while since I used blastp. Does it append to the file
given in '-out'? If not then you are overwriting it for every 24
sequences. Maybe you meant {#} instead?

>  sh -c 'dd bs=1 count=1 of=/tmp/pariINik.chr 2>/dev/null';  test ! -s
> "/tmp/pariINik.chr" && rm -f "/tmp/pariINik.chr" && exec true;  (cat
> /tmp/pariINik.chr; rm /tmp/pariINik.chr; cat - ) |
> (/afs/enea.it/software/bin/blaunch.sh cresco3x018.portici.enea.it exec perl
> -e
> \\\$ENV\\\{\\\"PARALLEL_PID\\\"\\\}=\\\"30669\\\"\\\;\\\$ENV\\\{\\\"PARALLEL_SEQ\\\"\\\}=\\\"687\\\"\\\;\\\$bashfunc\\\
> =\\\ \\\"\\\"\\\;@ARGV=\\\"blastp\\\ -evalue\\\ 1e-05\\\ -outfmt\\\ 6\\\
> -db\\\
> /gporq1_1M/usr/aprea/bio/solanum_melongena/analysis/orthomcl_00/goodProteins_first_0010000\\\
> -query\\\ -\\\ -out\\\
> /gporq1_1M/usr/aprea/bio/solanum_melongena/analysis/orthomcl_00/resultd_24\\\"\\\;\\\$SIG\\\{CHLD\\\}=sub\\\{\\\$done=1\\\;\\\}\\\;\\\$pid=fork\\\;unless\\\(\\\$pid\\\)\\\{setpgrp\\\;exec\\\$ENV\\\{SHELL\\\},\\\"-c\\\",\\\(\\\$bashfunc.\\\"@ARGV\\\"\\\)\\\;die\\\"exec:\\\$\\\!\\\\n\\\"\\\;\\\}do\\\{\\\$s=\\\$s\\\<1\\\?0.001+\\\$s\\\*1.03:\\\$s\\\;select\\\(undef,undef,undef,\\\$s\\\)\\\;\\\}until\\\(\\\$done\\\|\\\|getppid==1\\\)\\\;kill\\\(SIGHUP,-\\\$\\\{pid\\\}\\\)unless\\\$done\\\;wait\\\;exit\\\(\\\$\\\?\\\&127\\\?128+\\\(\\\$\\\?\\\&127\\\):1+\\\$\\\?\\\>\\\>8\\\););

-vv is really only useful for debugging: It is extremely hard to read
- even if you are the author of GNU Parallel.

I will highly recommend to use '-v' first and only resort to '-vv' if
'-v' shows what is expected.


/Ole