Editing
Lab IT
(section)
Jump to navigation
Jump to search
Warning:
You are not logged in. Your IP address will be publicly visible if you make any edits. If you
log in
or
create an account
, your edits will be attributed to your username, along with other benefits.
Anti-spam check. Do
not
fill this in!
==TSCC cluster== *UPDATE: Triton is expected to go down by the start of July 2013. Home files will be moved to the TSCC servers and /project will be remounted on the TSCC servers (warning: uid/gid might be changed on project drive!) * INFO on data transfer and storage on TSCC: http://rci.ucsd.edu/computing/storage/data-transfer.html * INFO on starting jobs on TSCC: http://rci.ucsd.edu/computing/jobs/index.html *The Zhang lab current has four "home" nodes each having 16 cores and 64Gb of RAM. This is the best option when dealing with a large amount of data, such sequencing reads from one or more HiSeq runs. To use these nodes, you will need your own accounts at SDSC TSCC cluster (tscc-login.sdsc.edu). Once you logging in your account, you can create job files, and submit them to the queue: [#PBS -A k4zhang-group] - this line may not need to be in job file to submit to home, condo, or glean. Then submit job to a specific queue with Submit to home queue: qsub -q home-k4zhang job_name Submit to condo queue: qsub -q condo job_name -W group_list=condo-group Submit to glean queue: qsub -q glean job_name -W group_list=condo-group Submit to hotel queue (pay-per-use): qsub -q hotel job_name Submit to pdafm queue (pay-per-use): qsub -q pdafm job_name * Our home nodes IDs are: '''tscc-2-33''', '''tscc-2-35''', '''tscc-2-37''', '''tscc-2-39''' * The cheapest nodes for running jobs are: '''glean''' (runs free, no time limit), '''home-k4zhang''' (run on purchased time, no time limit), '''condo''' (run on purchased time, 8 hrs limit). * The pay-as-you-go nodes (72 hours limit) for running jobs are: '''hotel''' (can specify GPU), and '''pdafm''' (max 512GB memory, specify less fewer processors/memory to get charged less) * '''Node recommendations''': High priority jobs should be run with '''home-k4zhang''' or '''condo'''. When running '''home-k4zhang''', we can only start a limited number of jobs. Also, run ''qstat'' to see that no one in our lab is using '''glean''' To start many high priority jobs which takes less than 8 hours to complete, use '''condo''' Intermediate to low priority can be run with '''glean''', there is unlimited time but may get kicked off at any time. Run '''glean''' on our nodes to lower the risks of getting kicked off, use any of the following lines to specify which of our nodes to run on: #PBS -l nodes=tscc-2-33:ppn=X #PBS -l nodes=tscc-2-35:ppn=X #PBS -l nodes=tscc-2-37:ppn=X #PBS -l nodes=tscc-2-39:ppn=X (X = number of processors, if not specified, the entire node might be used and we can't start more than one jobs on each node.) The only times we would get kicked off '''glean''' would be when someone in our lab starts a job with '''home-k4zhang''' or when '''condo''' is at full capacity (very unlikely!) * '''File storage and access''': Do not transfer large files with ''scp'', install and use ''bbftp'' and start a job to transfer files. ( Large files for ''multiple uses'' can be transferred to /oasis (it have faster read/write than on /project or /home)! Small or large files for ''one-time'' use can be transferred to the path provided by the TMPDIR environment variable (this will copy the file to a local scratch space) then immediately processed. Once the job finishes, any file in this local scratch space will be lost. Always copy results back to our lab server to avoid losing them. Avoid as much as possible reading and writing to /project or /home. Use /oasis instead. * '''Checking the output of a job while the job is still running:''' *#Use qstat -u [your_username_here] to get the job id: should look something like "XXXXXXX.tscc-mgr.local" where the 7 digits before ".tscc-mgr.local" are your job ID. *#Use checkjob [your_job_id_here] to get info about your job. The node your job is running on should be under "Allocated Nodes" in the format tscc-x-XX:YY where the two digits after the colon are the number of processors requested. *#ssh into the node your job is using "ssh tscc-x-XX" *#cd /var/spool/torque/spool *#There should be two files associated with your job id labeled something like "XXXXXXX.OU" and "XXXXXXX.ER". These are the stdout and stderr files, respectively and you can view them to check the output of your job. This is a convenient way to check job progress * ''' Installing perl modules locally''' **Use cpan to install perl modules. **First determine where you want your local Perl modules, ie /home/userID/perlmods. **From within CPAN shell, enter the following (2) commands: *# cpan[1]> o conf mbuildpl_arg "--install_base /home/userID/perlmods" *# cpan[2]> o conf makepl_arg "PREFIX=/home/userID/perlmods" ** If you want to save these settings permanently, type 'o conf commit' (without the quotes). Otherwise you need to run the above commands everytime you start CPAN. ** Now to install a module: *# cpan[3] install Statistics::Descriptive ** Next, set the PERL5LIB path by adding the following lines to ~/.bash_profile or ~/.bashrc *# export PERL5LIB=/home/userID/perlmods:$PERL5LIB ** Finally source the .bash_profile or .bashrc so you could use it immediately. *# source ~/.bash_profile
Summary:
Please note that all contributions to ZhangLabWiki may be edited, altered, or removed by other contributors. If you do not want your writing to be edited mercilessly, then do not submit it here.
You are also promising us that you wrote this yourself, or copied it from a public domain or similar free resource (see
ZhangLabWiki:Copyrights
for details).
Do not submit copyrighted work without permission!
Cancel
Editing help
(opens in new window)
Navigation menu
Personal tools
Not logged in
Talk
Contributions
Create account
Log in
Namespaces
Page
Discussion
English
Views
Read
Edit
View history
More
Search
Navigation
Main Page
Current events
Recent changes
Random page
Investigators
Matt Cai
Song Chen
Eric Chu
Dinh Diep
Elizabeth Duong
Shicheng Guo
Alan Fung
Daniel Jacobsen
Blue Lake
Huy Lam
Alice Li
Andrew Richards
Brandon Sos
Chris Wei
Yan Wu
Kun Zhang
Tools
What links here
Related changes
Special pages
Page information