Partitioning Mongrel Processes between Faspex and Cargo

Partition mongrels between handling Faspex UI requests and IBM Aspera Cargo requests to address performance issues.

When the number of Cargo clients attached to a Faspex cluster reaches a significant number, the performance of the Faspex Web UI can suffer due to resource contention with Cargo clients accessing the API.

To avoid this, tune the Apache configuration with separate sets of Mongrel processes dedicated to serving the Faspex web interface and API.

Note: The examples in the instructions below demonstrates Mongrel partitioning for 15 mongrel processes: 10 for Cargo and 5 for Faspex.

Run the following asctl command to set the proper total number of Mongrel processes:
```
# asctl faspex:mongrel_count number+of_mongrels
```
Note: Running the mongrel_count command overwrites and removes any modifications to the faspex.apache.linux.conf configuration file, including the changes described in the following steps.
Choose not to restart Apache and Faspex.

Open the following Faspex Apache configuration file in a text editor: /opt/aspera/faspex/config/faspex.apache.linux.conf and make the following changes

Add the <Proxy balancer://faspex_cargo_cluster> section.

For example:

...
#Proxy balancer section (create one for each ruby app cluster)
<Proxy balancer://faspex_cargo_cluster>
</Proxy>
<Proxy balancer://faspex_cluster>
  BalancerMember http://127.0.0.1:3000
  BalancerMember http://127.0.0.1:3001
  BalancerMember http://127.0.0.1:3002
  BalancerMember http://127.0.0.1:3003
  BalancerMember http://127.0.0.1:3004
  BalancerMember http://127.0.0.1:3005
  BalancerMember http://127.0.0.1:3006
  BalancerMember http://127.0.0.1:3007
  BalancerMember http://127.0.0.1:3008
  BalancerMember http://127.0.0.1:3009
  BalancerMember http://127.0.0.1:3010
  BalancerMember http://127.0.0.1:3011
  BalancerMember http://127.0.0.1:3012
  BalancerMember http://127.0.0.1:3013
  BalancerMember http://127.0.0.1:3014
</Proxy>
...

Distribute the BalancerMember entries under <Proxy balancer://faspex_cluster> between the two sections.

For example:

...
#Proxy balancer section (create one for each ruby app cluster)
<Proxy balancer://faspex_cargo_cluster>
  BalancerMember http://127.0.0.1:3000/aspera/faspex/inbox.atom
  BalancerMember http://127.0.0.1:3001/aspera/faspex/inbox.atom
  BalancerMember http://127.0.0.1:3002/aspera/faspex/inbox.atom
  BalancerMember http://127.0.0.1:3003/aspera/faspex/inbox.atom
  BalancerMember http://127.0.0.1:3004/aspera/faspex/inbox.atom
  BalancerMember http://127.0.0.1:3005/aspera/faspex/inbox.atom
  BalancerMember http://127.0.0.1:3006/aspera/faspex/inbox.atom
  BalancerMember http://127.0.0.1:3007/aspera/faspex/inbox.atom
  BalancerMember http://127.0.0.1:3008/aspera/faspex/inbox.atom
  BalancerMember http://127.0.0.1:3009/aspera/faspex/inbox.atom
</Proxy>
<Proxy balancer://faspex_cluster>
  BalancerMember http://127.0.0.1:3010
  BalancerMember http://127.0.0.1:3011
  BalancerMember http://127.0.0.1:3012
  BalancerMember http://127.0.0.1:3013
  BalancerMember http://127.0.0.1:3014
</Proxy>
...

Add ProxyPass /aspera/faspex/inbox.atom balancer://faspex_cargo_cluster to the proxy request section.

...
# send the proxy request
ProxyPass /aspera/faspex/inbox.atom balancer://faspex_cargo_cluster
ProxyPass /aspera/faspex balancer://faspex_cluster 
...

Restart Apache and Faspex services.

# asctl apache:restart
# asctl faspex:restart