run Trinity on a server

Question

[luigi.marongiu]

Hello,

I ran Trinity with this command

sudo Trinity --seqType fq --max_memory 10G --no_bowtie --no_version_check --left pairedA_c20lr1,pairedB_c20lr1 --right unpairedA_c20lr1,unpairedB_c20lr1  --CPU 10

but I got the present error:

 Running CMD: /usr/local/lib/Trinity/Inchworm/bin//inchworm --kmers jellyfish.kmers.fa --run_inchworm -K 25 -L 25 --monitor 1   --DS  --num_threads 6  --PARALLEL_IWORM  > /home/gigiux/SpiderOak Hive/LAB/Lab book/Nors/Sequences/trinity_out_dir/inchworm.K25.L25.DS.fa.tmp
sh: 1: cannot create /home/gigiux/SpiderOak: Is a directory
Error, cmd: /usr/local/lib/Trinity/Inchworm/bin//inchworm --kmers jellyfish.kmers.fa --run_inchworm -K 25 -L 25 --monitor 1   --DS  --num_threads 6  --PARALLEL_IWORM  > /home/gigiux/SpiderOak Hive/LAB/Lab book/Nors/Sequences/trinity_out_dir/inchworm.K25.L25.DS.fa.tmp 2>tmp.2633.stderr died with ret 512 at /usr/local/lib/Trinity/PerlLib/Pipeliner.pm line 102.
    Pipeliner::run(Pipeliner=HASH(0x28daa30)) called at /usr/bin/Trinity line 2054
    eval {...} called at /usr/bin/Trinity line 2049
    main::run_inchworm("/home/gigiux/SpiderOak Hive/LAB/Lab book/Nors/Sequences/trini"..., "both.fa", undef, "") called at /usr/bin/Trinity line 1424
    main::run_Trinity() called at /usr/bin/Trinity line 1206
    eval {...} called at /usr/bin/Trinity line 1205

and the most important bit:

If it indicates bad_alloc(), then Inchworm ran out of memory.  You'll need to either reduce the size of your data set or run Trinity on a server with more memory available.

It is really possible to run Trinity on a server? Would they work like those cloud folder services? Which ones would you recommend? Or it is just a syntax error of the command?

Thank you

Answer 1

[h.mon]

Read the error message carefully, from what you posted you do not have a memory problem (yet). The message you highlighted is just informative.

What do you mean by "run Trinity on a server"?

Running Trinity on a SpiderOak folder is a terrible, terrible idea for two reasons:

1. Trinity will create loads of data and you will run out of space for synchronization (Trinity will continue to run locally, though)
2. more important, there is a space on the folder name, and this may (almost certainly will) cause errors for command-line programs, and I suspect it is your current problem.

In addition, running anything not related to server maintenance as root / using sudo is also a bad idea.

[luigi.marongiu]

Thanks for the answer. Actually you are spotting some of the troubles I got with the command. I use SpiderOak because it is a shared folder that I can access at home and at work; it behaves as a normal folder on the PC so as long as I have hard drive space, there should not be a problem, I clean up and export data afterwards and keep aside the sync for the moment. The problem is that Trinity creates new folders with the content of SpiderOak once again and this eats a lot of space. As for the space (point 2), does this refers to /Spider<whitespace>Oak/ and Lab<whitespace>book? Running with sudo comes from the fact that without it I get the error for both threads:

readcount died with ret 256 at /usr/bin/Trinity line 2183

[h.mon]

Yes, I was referring to /Spider<whitespace>Oak/ and Lab<whitespace>book. Spaces on file or folder names are generally a bad idea, they will often cause command line (and sometimes GUI) programs to crash. Regarding your sudo problems, you have to ask the systems admin to fix your permissions - but this is not a bioinformatics question.

I know SpiderOak, and I know Trinity will easily eat up all your space there. And worst, synchronizing while running the analysis (and generating thousands of files) will increase IO and probably increase assembly time.