FTS Server Upgrade 1.4 to 1.5 for LCG service

This assumes you are running the latest 1.4 FTS version and describes how to upgrade the service to FTS 1.5. It is not recommended to upgrade from any other release, nor to apply cumulative upgrades from the earlier releases (instead make a clean install FtsServerInstall15).

Read through all the steps first.

Changes from release 1.4 are described in FtsChangesFrom13To14.

Requirements

Are the same as for your current 1.4 version, see FtsServerInstall14 and FtsServerInstall15; the deployment model is the same.

It is recommended to split the FTS web-service onto a different box from the FTA transfer agents for scalability.

Make sure you know the Grid dependencies

The Grid dependencies are the same as in 1.4, see FtsServerInstall14 and FtsServerInstall15.

You should ensure that all the SRM you plan to contact are published correctly in EGEE.BDII.

Upgrade recommendation

For the upgrade from 1.4, the recommendation is the completely re-install the nodes from scratch.

There are several tricky problems due to dependency changes which will prevent a clean and easy software upgrade. Additionally, the configuration mechanism has been changed to YAIM which is a significant change.

There is a script to upgrade the DB schema so the current job information and queues will be maintained.

When can I not upgrade?

Check your channel names. They MUST contain ONLY:

  • alphanumeric upper-case characters or a dash -.

If they do not fit the pattern, you should drop the schema and proceed with a fresh install from FtsServerInstall15.

As alternative, you could use the FTS admin pack described in FtsAdminToolsPackageLoading15, but for the moment the fresh DB install is the recommended solution.

Channel Definitions

Please note also that in case the channel definitions contain site names with lowercase letter, you have to modify them in order to have all these names uppercase. You can find the explanation and the description here.

Set nodes to maintenance

CERN-specific:

  • Set nodes to SMS state maintenance
  • Ensure that the TOMCAT_WRONG alarm and actuator is disabled.
  • Inform the operators.

Or do the equivalent on your site to signal a service intervention.

Drain the service

Set all transfer channels inactive - this will drain the network of active jobs and ensure their status is comitted fully to the database.

for i in `glite-transfer-channel-list`; do glite-transfer-channel-set $i Inactive; done

Go for a coffee

If all channels are working properly, it should take 5 to 15 minutes to drain the channels. If there are problems with some of the SRM, you may have to wait upto 30 minutes for bad transfers to time out.

Stop the servers

Stop the agents and tomcat servers on all machines.

service tomcat5 stop

/opt/glite/etc/config/scripts/glite-file-transfer-agents-config.py --stop

Make sure they are stopped:

ps aux | grep j2
ps aux | grep glite

should produce nothing except the grep.

Remove any other open sessions you may have to the DB account.

Archive the logfiles

See the FtsServerAdmin14 for details of logfiles and archive those that you wish to archive.

Make a note of server specific settings

FTS web-service

Look in the files:

/opt/glite/etc/config/templates/glite-file-transfer-service.cfg.xml

and make a note of any settings you had changed, noting the name and the value. Note in particular, the DB connection URL and the username and password.

FTA agents

/opt/glite/etc/config/templates/glite-file-transfer-agents.cfg.xml

In particular, note any parameters inside the instance XML blocks. Note the name of which agent it applies to, the parameter name that was changed and the value that it was changed to. There should be relatively few changes, since most (if not all) of the agents' parameters will the default.

Make a note of which agents you have, their names, and whether thay are running urlcopy, srmcopy or are VO agents.

Wipe the machine(s)

Install a fresh SLC3 installation as per the LCG instructions.

Proceed with the fresh install instructions

The instructions are given in FtsServerInstall15. Use the DB connection params you saved from the previous config.

Note a couple of things below.

FTS web-service

The configuration of the web-service will stop with the error:

Database schema is using the old 2.1.0 version. Please upgrade using the command:

    sqlplus user/pass@DBname @/opt/glite/etc/glite-data-transfer-fts/schema/oracle/oracle-upgrade_2.1.0-2.2.0.sql

NOTE! This operation is irreversible!

Run the suggested command and it will upgrade the schema to the current 2.2.0 version. Be sure that there are no agents or web-services before running this script. As the docu says, the operation on the schema is not reversible.

The upgrade may take a while (depending on how many entries you have in the DB), so consider going for a coffee.

Rerun the configuration script for FTS web-service and proceed with the rest of the instructions.

FTA agent servers

Define all the agents you noted previously in the YAIM site-info.def file.

The 'name' attribute of the instance block is what should be used to identify the agent in YAIM.


Last edit: FlaviaDonno on 2007-02-02 - 10:03

Number of topics: 1

Maintainer: GavinMcCance


Edit | Attach | Watch | Print version | History: r5 < r4 < r3 < r2 < r1 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r5 - 2007-03-23 - SteveTraylen
 
    • Cern Search Icon Cern Search
    • TWiki Search Icon TWiki Search
    • Google Search Icon Google Search

    LCG All webs login

This site is powered by the TWiki collaboration platform Powered by PerlCopyright & 2008-2019 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback