Install the Frontier Squid HTTP Caching Proxy¶
Frontier Squid is a distribution of the well-known squid HTTP caching proxy software that is optimized for use with applications on the Worldwide LHC Computing Grid (WLCG). It has many advantages over regular squid for common distributed computing applications, especially Frontier and CVMFS. The OSG distribution of frontier-squid is a straight rebuild of the upstream frontier-squid package for the convenience of OSG users.
This document is intended for System Administrators who are installing
frontier-squid, the OSG distribution of the Frontier Squid software.
Frontier Squid Is Recommended¶
OSG recommends that all sites run a caching proxy for HTTP and HTTPS to help reduce bandwidth and improve throughput. To that end, Compute Element (CE) installations include Frontier Squid automatically. We encourage all sites to configure and use this service, as described below.
For large sites that expect heavy load on the proxy, it is best to run the proxy on its own host. If you are unsure if your site qualifies, we recommend initially running the proxy on your CE host and monitoring its bandwidth. If the network usage regularly peaks at over one third of the bandwidth capacity, move the proxy to a new host.
Before starting the installation process, consider the following points (consulting the Reference section below as needed):
- User IDs: If it does not exist already, the installation will create the
- Network ports: Clients within your cluster (e.g., OSG user jobs) will communicate with Frontier Squid on port 3128 (TCP). Additionally, central infrastructure will monitor Frontier Squid through port 3401 (UDP); see this section for more details.
- Host choice: If you will be supporting the Frontier application at your site, review the upstream documentation to determine how to size your equipment.
As with all OSG software installations, there are some one-time (per host) steps to prepare in advance:
- Ensure the host has a supported operating system
- Obtain root access to the host
- Prepare the required Yum repositories
Installing Frontier Squid¶
To install Frontier Squid, make sure that your host is up to date before installing the required packages:
Clean yum cache:
[email protected] # yum clean all --enablerepo=*
[email protected] # yum update
This command will update all packages
Install Frontier Squid:
[email protected] # yum install frontier-squid
Configuring Frontier Squid¶
Configuring the Frontier Squid Service¶
To configure the Frontier Squid service itself:
- Follow the Configuration section of the upstream Frontier Squid documentation.
- Enable, start, and test the service (as described below).
- Register the squid (also as described below).
An important difference between the standard Squid software and
the Frontier Squid variant is that Frontier Squid changes are in
/etc/squid/customize.sh instead of
Configuring the OSG CE¶
To configure the OSG Compute Entrypoint (CE) to know about your Frontier Squid service:
On your CE host (which may be different than your Frontier Squid host), edit
- Make sure that
enabledis set to
locationto the hostname and port of your Frontier Squid service (e.g.,
- Leave the other settings at
DEFAULTunless you have specific reasons to change them
- Make sure that
osg-configure -cto propagate the changes on your CE.
You may want to finish other CE configuration tasks before running
osg-configure. Just be sure to run it once before starting CE
Start the frontier-squid service and enable it to start at boot time. As a reminder, here are common service commands (all run as
|To...||Run the command...|
|Start the service||
|Stop the service||
|Enable the service to start on boot||
|Disable the service from starting on boot||
Validating Frontier Squid¶
As any user on another computer, do the following (where
<MY.SQUID.HOST.EDU> is the fully qualified domain name of your
[email protected] $ export http_proxy=http://`<MY.SQUID.HOST.EDU>`:3128 [email protected] $ wget -qdO/dev/null http://frontier.cern.ch 2>&1|grep X-Cache X-Cache: MISS from `<MY.SQUID.HOST.EDU>` [email protected] $ wget -qdO/dev/null http://frontier.cern.ch 2>&1|grep X-Cache X-Cache: HIT from `<MY.SQUID.HOST.EDU>`
If the grep doesn't print anything, try removing it from the pipeline to see if errors are obvious. If the second try says MISS again, something is probably wrong with the squid cache writes. Look at the squid access.log file to try to see what's wrong.
If your squid will be supporting the Frontier application, it is also good to do the test in the upstream documentation Testing the installation section.
Registering Frontier Squid¶
To register your Frontier Squid host, follow the general registration instructions here with the following Frontier Squid-specific details. Alternatively, contact us for assistance with the registration process.
Squid:section to the
Services:list, with any relevant fields for that service. This is a partial example:
... FQDN: <FULLY QUALIFIED DOMAIN NAME> Services: Squid: Description: Generic squid service ...
<FULLY QUALIFIED DOMAIN NAME>with your Frontier Squid server's DNS entry or in the case of multiple Frontier Squid servers for a single resource, the round-robin DNS entry.
See the BNL_ATLAS_Frontier_Squid for a complete example.
Normally registered squids will be monitored by WLCG. This is strongly recommended even for non-WLCG sites so operations experts can help with diagnosing problems. However, if a site declines monitoring, that can be indicated by setting
Monitored: falsein a
Description:. Registration is still important for the sake of excluding squids from worker node failover monitors. The default if
Monitored:is not set is
If you set Monitored to true, also enable monitoring as described in the upstream documentation on enabling monitoring.
A few hours after a squid is registered and marked
Active (and not
verify that it is monitored by WLCG.
The frontier-squid installation will create one user account unless it already exists.
||Reduced privilege user that the squid process runs under. Set the default gid of the "squid" user to be a group that is also called "squid".|
The package can instead use another user name of your choice if you create a configuration file before installation. Details are in the upstream documentation Preparation section.
Open the following ports on your Frontier Squid hosts:
|3128||tcp||✓||Also limited in squid ACLs. Should be limited to access from your worker nodes|
|3401||udp||✓||Also limited in squid ACLs. Should be limited to public monitoring server addresses|
The addresses of the WLCG monitoring servers for use in firewalls are listed in the upstream documentation Enabling monitoring section.
Frontier Squid Log Files¶
Log file contents are explained in the upstream documentation Log file contents section.