To submit your job you will need to have created your job file and have your data files ready.
/pool
(/pool/genomics
for genomics users) or /data
.mkdir
from the command line. Log into Hydra in the terminal and navigate to the directory with your data files.
Throughout this tutorial replace username
your Hydra username, the name you log into Hydra with, typically the same as your SI email account name (all lower case).
[username@hydra-login01 ~]$ cd /pool/genomics/username/raxmltest/ [username@hydra-login01 raxmltest]$ |
Use ls
to confirm your job and data files are present and use cat
(or another text file viewer like less
) to inspect your job file.
[username@hydra-login01 raxmltest]$ ls input.fasta raxmltest.job [username@hydra-login-1 raxmltest]$ cat raxmltest.job # /bin/sh # ----------------Parameters---------------------- # #$ -S /bin/sh #$ -q mThC.q #$ -l mres=2G,h_data=2G,h_vmem=2G #$ -cwd #$ -j y #$ -N raxmltest #$ -o raxmltest.log #$ -m abe #$ -M user@si.edu # # ----------------Modules------------------------- # module load bioinformatics/raxml/latest # # ----------------Your Commands------------------- # # echo + `date` job $JOB_NAME started in $QUEUE with jobID=$JOB_ID on $HOSTNAME # raxmlHPC-SSE3 -f a -p 12345 -s input.fasta -x 12345 -N 100 -m GTRCAT -n testrun # echo = `date` job $JOB_NAME done |
Submit your job to the cluster with the qsub
command.
[username@hydra-login01 raxmltest]$ qsub raxmltest.job Your job 825184 ("raxmltest") has been submitted |
Check your job status with the qstat
command and ls
Errors in the job file will often cause the job to fail quickly, so check qstat
during the first few minutes of a job run to confirm that it was submitted successfully. If nothing is displayed then you have no current jobs.
[username@hydra-login01 raxmltest]$ qstat -u username job-ID prior name user state submit/start at queue slots ja-task-ID ----------------------------------------------------------------------------------------------------------------- 825184 0.55500 raxmltest username r 09/22/2015 10:18:01 mThC.q@compute-81-01.cm.clust 1 [username@hydra-login01 raxmltest]$ ls input.fasta RAxML_bootstrap.testrun RAxML_info.testrun raxmltest.job raxmltest.log |
You can now logout of Hydra with the exit
command and log back in later to check your job progress
When your job is complete the job will no longer appear in qstat and you will receive an email if you configured your job file to send email notifications. Inspect the output files to confirm that job was successful.
[username@hydra-login01 raxmltest]$ qstat -u username [username@hydra-login01 raxmltest]$ tail raxmltest.log Best-scoring ML tree written to: /pool/genomics/username/raxmltest/RAxML_bestTree.testrun Best-scoring ML tree with support values written to: /pool/genomics/username/raxmltest/RAxML_bipartitions.testrun Best-scoring ML tree with support values as branch labels written to: /pool/genomics/username/raxmltest/RAxML_bipartitionsBranchLabels.testrun Overall execution time for full ML analysis: 149.352777 secs or 0.041487 hours or 0.001729 days = Tue Sep 22 10:20:30 EDT 2015 job raxmltest done |
Transfer the output files to your computer or continue with further analysis of the dataset on Hydra.
It is recommended to compress your output before transferring to your local computer
[username@hydra-login01 raxmltest]$ cd .. [username@hydra-login01 username]$ tar -czvf raxmltest.tar.gz raxmltest raxmltest/ raxmltest/raxmltest.job raxmltest/input.fasta raxmltest/raxmltest.log raxmltest/RAxML_info.testrun raxmltest/RAxML_bootstrap.testrun raxmltest/RAxML_bestTree.testrun raxmltest/RAxML_bipartitions.testrun raxmltest/RAxML_bipartitionsBranchLabels.testrun |
Storage on Hydra is not for long-term data storage. Please transfer and remove your data files and output when your analyses are complete.
Last updated SGK