How to run DMT Test via SeqInsert

In an engineering run, the actual dmt triggers are sent via a ligotools program SeqInsert via an http like process. This test exercises the method to make sure ldas is compatible with actual dmt process; the test objectives are the same as simulated DMT:

The LDAS APIs affected by this test are:

Setup

Choose a directory to install the tests scripts.

Running the test

The test should be run from your local platform, and run in the scripts directory. Before running the test, clear out any previous dmt data in the database you are going to do insertion by running the delall.sql script from ldas installation: e.g.


			db2 'connect to my_db'
			db2 -tf /ldas/doc/db2/doc/text/delall.sql 
			cd <scripts directory>
			tclsh insertTrigs cit ldas_tst ../triggers2 >& cit.log &
		
Usage: tclsh insertTrigs [options] >& insertTrigs.log &

Option Description
<site> ldas site [ dev | test | lho | llo | mit ]

or <mycomputer>:10001

<dsname> database name at a site e.g. cit_test
<dmt triggers directory> directory holding dmt triggers xml files e.g. /home/john/triggers2

Test Verification:

The test should be monitored initially to make sure data is flowing in a steady pace and that the insertions are successful impling that data is inserted in the correct sequence.

  1. Monitoring dmt

    All output from script SeqInsert is captured to stdout which can be redirected to a file at startup e.g. cit.log.

    A putMetaData user command is generated for each of the dmt files and sent to the managerAPI. The insertions can be monitored via cmonClient for the number of jobs and the database insertions for the specified database.

  2. Analysis

    When all the xmls are sent to LDAS, the logs can be extracted via cmonClient Log page; enter the gpstime ranges recorded in insertTrigs.log for logs in managerAPI, ligolwAPI and metadataAPI. Validate the following is true for successful insertions:

    If there is a discrepancy search the logs for red balls to account for missing records.

    Also examine the ldas database web page for the number of processes, segments and gds_triggers inserted to make sure they match the expected number.

    If all the process and trigger data are inserted, the test passes. If not, the test fails and the errors in the logs should account for all data received at LDAS.