HTCondor-CE 3 Documentation
This documentation is for HTCondor-CE 3. If you need documentation for HTCondor-CE 4, please consult https://htcondor-ce.readthedocs.io/en/latest/
Installing an HTCondor-CE¶
If you are installing an HTCondor-CE for the Open Science Grid (OSG), consult the OSG-specific documentation.
The HTCondor-CE software is a job gateway based on HTCondor for Compute Elements (CE) belonging to a computing grid (e.g. European Grid Infrastructure, Open Science Grid). As such, HTCondor-CE serves as an entry point for incoming grid jobs — it handles authorization and delegation of jobs to a grid site's local batch system. See the overview page for more details on the features and architecture of HTCondor-CE.
Use this page to learn how to install, configure, run, test, and troubleshoot HTCondor-CE from the HTCondor Yum repositories.
Before starting the installation process, consider the following points (consulting the reference page as necessary):
- User IDs: If they do not exist already, the installation will create the
condorLinux user (UID 4716)
- SSL certificate: The HTCondor-CE service uses a host certificate at
/etc/grid-security/hostcert.pemand an accompanying key at
- DNS entries: Forward and reverse DNS must resolve for the HTCondor-CE host
- Network ports: The pilot factories must be able to contact your HTCondor-CE service on port 9619 (TCP)
- Submit host: HTCondor-CE should be installed on a host that already has the ability to submit jobs into your local cluster running supported batch system software (Grid Engine, HTCondor, LSF, PBS/Torque, Slurm)
- File Systems: Non-HTCondor batch systems require a shared file system between the HTCondor-CE host and the batch system worker nodes.
There are some one-time (per host) steps to prepare in advance:
- Ensure the host has a supported operating system (Red Hat Enterprise Linux variant 7)
- Obtain root access to the host
- Prepare the EPEL and HTCondor Yum repositories
- Install CA certificates and VO data into
HTCondor-CE must be installed on a host that is configured to submit jobs to your batch system. The details of this setup is site-specific by nature and therefore beyond the scope of this document.
Clean yum cache:
[email protected] # yum clean all --enablerepo=*
[email protected] # yum update
This command will update all packages
fetch-crlpackage, available from the EPEL repositories.
[email protected] # yum install fetch-crl
Select the appropriate convenience RPM:
If your batch system is... Then use the following package... Grid Engine
Install the CE software:
[email protected] # yum install <PACKAGE>
<PACKAGE>is the package you selected in the above step.
There are a few required configuration steps to connect HTCondor-CE with your batch system and authentication method. For more advanced configuration, see the section on optional configurations.
To authenticate job submission from external users and VOs, HTCondor-CE can be configured to use a built-in mapfile or to make Globus callouts to an external service like Argus or LCMAPS. THe former option is simpler but the latter option may be preferred if your grid supports it or your site already runs such a service.
The built-in mapfile is a
unified HTCondor mapfile
This file is parsed in line-by-line order and HTCondor-CE will use the first line that matches.
Therefore, mappings should be added to the top of the file.
condor_mapfile.rpmnew files may be generated upon HTCondor-CE version updates and they should be merged into
To configure authorization for users submitting jobs with an X.509 proxy certificate to your HTCondor-CE, add lines of the following format:
GSI "^<DISTINGUISHED NAME>$" <USERNAME>@users.htcondor.org
<DISTINGUISHED NAME> (escaping any '/' with '\/') and
<USERNAME> with the distinguished name of the
incoming certificate and the unix account under which the job should run, respectively.
VOMS attributes of incoming X.509 proxy certificates can also be used for mapping:
GSI "<DISTINGUISHED NAME>,<VOMS FQAN 1>,<VOMS FQAN 2>,...,<VOMSFQAN N>" <USERNAME>@users.htcondor.org
<DISTINGUISHED NAME> (escaping any '/' with '\/'),
<VOMSFQAN> fields, and
<USERNAME> with the
distinguished name of the incoming certificate, the VOMS roles and groups, and the unix account under which the job
should run, respectively.
Additionally, you can use regular expressions for mapping certificate and VOMS attribute credentials.
For example, to map any certificate from the
GLOW VO with the
htpc role to the
glow user, add the following line:
GSI ".*,\/GLOW\/Role=htpc.*" [email protected]
You should only add mappings to the mapfile. Do not remove any of the default mappings:
To use a Globus callout to a service like LCMAPS or Argus, you will need to have the relevant library installed as well as the following HTCondor-CE configuration:
Add the following line to the top of
GSI (.*) GSS_ASSIST_GRIDMAP
/etc/grid-security/gsi-authz.confwith the following content:
globus_mapping liblcas_lcmaps_gt4_mapping.so lcmaps_callout
globus_mapping /usr/lib64/libgsi_pep_callout.so argus_pep_callout
Configuring the batch system¶
Before HTCondor-CE can submit jobs to your local batch system, it has to be configured to do so. The configuration will differ depending on if your local batch system is HTCondor or one of the other supported batch systems. Choose the section corresponding to your batch system below.
HTCondor batch systems¶
To configure HTCondor-CE for an HTCondor batch system, set
JOB_ROUTER_SCHEDD2_POOL to your site's central manager host
JOB_ROUTER_SCHEDD2_POOL = cm.chtc.wisc.edu:9618
JOB_ROUTER_SCHEDD2_SPOOL to the location of the local batch
SPOOL directory on the CE host if it
is different than the default location (
Non-HTCondor batch systems¶
Configuring the BLAHP¶
HTCondor-CE uses the Batch Language ASCII Helper Protocol (BLAHP) to submit and track jobs to non-HTCondor batch systems.
To work with the HTCondor-CE, modify
/usr/libexec/condor/glite/etc/batch_gahp.config using the following steps:
Disable BLAHP handling of certificate proxies:
blah_disable_wn_proxy_renewal=yes blah_delegate_renewed_proxies=no blah_disable_limited_proxy=yes
(Optional) If your batch system tools are installed in a non-standard location (i.e., outside of
/usr/bin/), set the corresponding
*_binpathvariable to the directory containing your batch system tools:
If your batch system is... Then change the following configuration variable... LSF
For example, if your Slurm binaries (e.g.
sbatch) exist in
/opt/slurm/bin, you would set the following:
Sharing the SPOOL directory¶
Non-HTCondor batch systems require a shared file system configuration to support file transfer from the HTCondor-CE to
your site's worker nodes.
The current recommendation is to run a dedicated NFS server on the CE host.
In this setup, HTCondor-CE writes to the local spool directory, the NFS server shares the directory, and each worker
node mounts the directory in the same location as on the CE.
For example, if your spool directory is
/var/lib/condor-ce (the default), you must mount the shared directory to
/var/lib/condor-ce on the worker nodes.
If you choose not to host the NFS server on your CE, you will need to turn off root squash so that the HTCondor-CE daemons can write to the spool directory.
You can control the value of the spool directory by setting
this file if it doesn't exist).
For example, the following sets the
SPOOL directory to
SPOOL = /home/condor
The shared spool directory must be readable and writeable by the
condor user for HTCondor-CE to function correctly.
The following configuration steps are optional and will not be required for all sites. If you do not need any of the following special configurations, skip to the section on next steps.
Configuring for multiple network interfaces¶
If you have multiple network interfaces with different hostnames, the HTCondor-CE daemons need to know which hostname
and interface to use when communicating with each other.
NETWORK_INTERFACE to the hostname and IP address of your public interface, respectively, in
/etc/condor-ce/config.d/99-local.conf directory with the line:
NETWORK_HOSTNAME = condorce.example.com NETWORK_INTERFACE = 127.0.0.1
condorce.example.com text with your public interface’s hostname and
127.0.0.1 with your public interface’s
Enabling the monitoring web interface¶
The HTCondor-CE View is an optional web interface to the status of your CE. To run the HTCondor-CE View, install the appropriate package and set the relevant configuration.
Begin by installing the
[email protected] # yum install htcondor-ce-view
Next, uncomment the
DAEMON_LISTconfiguration located at
DAEMON_LIST = $(DAEMON_LIST), CEVIEW, GANGLIAD, SCHEDD
Verify the service by entering your CE's hostname into your web browser
The website is served on port 80 by default.
To change this default, edit the value of
Uploading accounting records to APEL¶
For sites outside of the OSG that need to upload the APEL accounting records, HTCondor-CE supports uploading batch and blah APEL records for HTCondor batch systems:
Install the HTCondor-CE APEL package on your CE host:
[email protected] # yum install htcondor-ce-apel
On each worker node, set the appropriate scaling factor in the HTCondor configuration (i.e.
/etc/condor/config.d/) and advertise it in the startd ad:
ApelScaling = <SCALING FACTOR> # For example, 1.062 STARTD_ATTRS = $(STARTD_ATTRS) ApelScaling
Configure the APEL parser, client, and SSM
- Records are written to
APEL_OUTPUT_DIRin the HTCondor-CE configuration (default:
- Batch and blah record filenames are prefixed
- Records are written to
Enabling BDII integration¶
HTCondor-CE supports reporting BDII information for all HTCondor-CE endpoints and batch information for an HTCondor batch system. To make this information available, perform the following instructions on your site BDII host.
Install the HTCondor-CE BDII package:
[email protected] # yum install htcondor-ce-bdii
Configure HTCondor (
/etc/condor/config.d/) on your site BDII host to point to your central manager:
CONDOR_HOST = <CENTRAL MANAGER>
<CENTRAL MANAGER>with the hostname of your HTCondor central manager
Configure BDII static information by modifying
Additionally, install the HTCondor-CE BDII package on each of your HTCondor-CE hosts:
[email protected] # yum install htcondor-ce-bdii
At this point, you should have an installation of HTCondor-CE that will forward grid jobs into your site's batch system unchanged. If you need to transform incoming grid jobs (e.g. by setting a partition, queue, or accounting group), configure the HTCondor-CE Job Router. Otherwise, continue to the this document to start the relevant services and verify your installation.
If you have any questions or issues with the installation process, please contact us for assistance,