ELabs Cluster Operations Power Up Checklist Reporting Bugs Starting and stopping worker nodes Database Servers Status Starting Stopping Deploying latest SVN

ELabs Cluster Operations

This page describes proceedures for managing the cluster, installing software, and the overall configuration of the cluster. See also the Cluster Configuration page for configuration information. Please keep the distinction between the two in mind when adding or editing content.

Real-time status of each machine is available at http://www12.i2u2.org:9080/elab/cosmic/monitor/index.jsp (requires SVG support)

Power Up Checklist

Whenever one or more machines is rebooted or powered up from maintenance, someone should run the Cluster Power-Up Checklist to verify that everything is functional.

Reporting Bugs

• Users (teachers and students) are now encouraged to report bugs via http://www13.i2u2.org/HelpDeskRequest.php, which sends a copy of the report to the list help@i2u2.org as well as to Eric and Mihael. It also posts a copy in the Help Desk forums
• If the request is really a bug, then enter it into our Bugzilla at http://bugzilla.mcs.anl.gov/i2u2/ The priorities are ordered from 1 = high, 5 = low as:
  • Priority 1: Some functionality breaks, most of the time the JSP scripts get invoked with bad inputs. All handled by Mihael; waiting for tester;s fedback
  • Priority 2: Cleanup, factoring, testing, load distributing issues. Some bing worked on (Liz, Ben) some pending (Tibi)
  • Priority 3: General improvements to the analysis, to the user interface, to usability. Verify the CMS elab from the scientific p.o.v
  • Priority 4: More further usability/functionality requirements

Starting and stopping worker nodes

Each worker node has /etc/init.d/i2u2tomcat which can be used to start or stop the server. There is also a restart option which is unreliable and should probably be removed - don't use it.

If you are authorised:

sudo service tomcat6 restart
sudo service apache2 restart

Database Servers

Both postgres and mysqld are run on the database server machine, data1, and both are supposed to be started at boot time. The examples below use posgres, but also apply to mysqld -- just substitute the names.

Status

To check the status of a daemon:

service postgres status

Starting

To start the database daemon by hand:

service postgresql-8.4 stop

Stopping

To stop the database daemon by hand:

service postgresql-8.4 stop

Deploying latest SVN

This is old! We now use SVN.

As yourself:

cd ~quarkcat
./deploy-from-svn branches/2.0