FTS service procedures: new node install

These are internal FTS procedures for the FIO managed CERN-PROD FTS services (FTS-T0-EXPORT and FTS-T2-SERVICE).

What is it?

After a new node install, to check the node is running.

When to use it?

After a new node install, to check the node is running.

Prerequisites

It is assumed that you have entered the new node profile into CDB to specify what the machine is (which service it is for and what its role is). It is assumed that you have set the necessary yaim confuiguration to include the new machine and to specify which agents reside on it.

The CDB setup for FTS is described in:

• FtsTier0CDBConfiguration20 for the SLC3-based services
• FtsTier0CDBConfiguration21 for the SLC4-based services

The following describes the steps necessary to get the node into production.

Web-servce node

After install, the web-service node ("FTS") should come up running.

You should check that both tomcat and BDII are running. See FtsServerAdmin20.

Set the SMS state to production (its default state).

Agent node

After a new install, and agent node may have come up badly due to a missing services.xml cache file which is generated daily from BDII (it's a cache of the BDII GlueService data).

Run the cache cron by hand:

/etc/cron.daily/glite-sd2cache-cron

Restart the agents:

service transfer-agents restart

This should be automated.

Check the agent processes are running. See FtsServerAdmin20.

Set the SMS state to production (its default state).

FTS monitoring node

This node should come up running with httpd.

Check httpd is running.

Set the SMS state to production (its default state).