FTS web-service software rolling upgrade procedure for Release 2.0

What is it?

This is the procedure for performing a rolling upgrade to the FTS web-service nodes.

The rolling upgrade assumes you have more than one FTS web-service and have a load-balanced production alias.

If you do not, the downtime is around 1 minute, depending on how fast you type.

When to use it?

When you want to upgrade the FTS web-service or change its configuration files without causing service downtime.

Procedure

We use in the example the three-node load-balanced cluster at CERN. The SMS controls are part of the CERN CC environment; substitute appropriately for your site. In this example the hosts are fts101, fts114 and fts115 but for the exact nodes check the nodes table on FtsTier0Deployment.

Mark one of the nodes for maintenance and drop it from the alias

sms set maintenance other "FTS upgrade" fts101

Wait for the node to drop out of the load-balanced alias.

host prod-fts-ws.cern.ch

If you are not using any load balancing you will just have to stop tomcat.

Upgrade the software / config files

Upgrade the web-application resource file RPM glite-data-transfer-fts and/or make the necessary YAIM configuration changes:

spma_wrapper.sh

[ edit the YAIM configuration files ]

Perform the FtsProcedureTomcatClean20 procedure to clean Tomcat.

Rerun the YAIM config tool:

/opt/glite/yaim/scripts/configure_node \
    /opt/glite/yaim/etc/site-info.def FTS2

This will restart the web-service. In the case of CERN follow the instructions for CDB configuration of yaim.

Re-enable the node

sms set production none "FTS upgrade done" fts101

Wait for the host to come back into the load-balanced alias before doing the other node:

host prod-fts-ws.cern.ch

Repeat for the other nodes

For CERN-PROD, this is fts114 and fts115.

Maintainer: GavinMcCance