Quick start¶

Just enough to get you going quickly. Read through the other pages in the list below the Orchestration topic for more details.

Warning

Launching HTCondor jobs through runOrca.py is done on the machine lsst-dev. Log in there to execute your jobs.

Create a $HOME/.lsst directory with permissions 700.

mkdir $HOME/.lsst
chmod 700 $HOME/.lsst

Create a db-auth.py configuration file with your MySQL host, user, password and mysql port information. This from must have permissions 600.
```
config.database.authInfo["auth1"].host = "lsst10.ncsa.illinois.edu"
config.database.authInfo["auth1"].user = "juser"
config.database.authInfo["auth1"].password = "funkystuff"
config.database.authInfo["auth1"].port = 3306
```
```
chmod 600 $HOME/.lsst/db-auth.py
```
Note

If you don’t already have a MySQL user account on the lsst10.ncsa.illinois.edu MySQL server, you’ll need to request one by filing a JIRA ticket in the IT Helpdesk Support (IHS) project.

Include the following information:
```
- Your name
- Institution and LSST affiliation
- Your email address
- Desired account name
```
You’ll also have to create a db-auth.paf file, because the pex_persistence package wasn’t updated to use Config. This file also belongs in $HOME/.lsst, and requires permissions 600.
```
database: {
    authInfo: {
        host: lsst-db.ncsa.illinois.edu
        port: 3306
        user: <user>
        password: <password>
    }
}
```
Create a HTCondor configuration file in $HOME/.lsst/condor-info.py
```
root.platform["lsst"].user.name = "juser"
root.platform["lsst"].user.home = "/lsst/home/juser"
```
Yes, this looks like something you shouldn’t have to specify. This is done for a consistent interface between platforms, since not all systems have consistent user names and home directories between sites or execution machines.
Create a directory named $HOME/condor_scratch
```
mkdir $HOME/condor_scratch
```

Setup ctrl_execute and ctrl_platform_lsst

setup ctrl_execute
setup ctrl_platform_lsst

Execute runOrca.py with the command you want to run

runOrca.py -p lsst -c "processCcdSdss.py sdss /lsst7/stripe82/dr7-coadds/v5/run0/jbosch_2012_0710_192216/input --output ./output" -i $HOME/short.input -e /lsst/DC3/stacks/gcc445-RH6/default

When you run this command, you’ll be told which identifier was created for the run. In the example, this jbosch_2012_0710_192216. The command will create directories under $HOME/condor_scratch and (for the LSST platform as it’s configured in lsst_ctrl_platform) under /lsst/DC3root, both named jbosch_2012_0710_192216.

This command says to run the command

processCcdSdss.py sdss /lsst7/stripe82/dr7-coadds/v5/run0/jbosch_2012_0710_192216/input --output ./output

using ids from the file $HOME/short.input executing out of an LSST stack located in /lsst/DC3/stacks/gcc445-RH6/default on the lsst platform.

Warning

This takes the user’s current EUPS environment and replicates it on the remote systems where to code is executed. The stack you’re pointing to on the remote system must have all the packages available to it that you do when you launch the command.

The results from the HTCondor output are for this run are in $HOME/condor_scratch/jbosch_2012_0710_192216 and the command output is under /lsst/DC3root/jbosch_2012_0710_192216.