30 | | Repeated testing with various database dumps from the live CaTissue v1.2 system are possible by dropping the database and creating it again, then loading the dump before running 'ant upgrade_db' from the catissue home directory. The csm_migrate_user fix above will need to be re-run each time. However, 'ant upgrade_db' did not successfully upgrade the database. |
| 30 | Repeated testing with various database dumps from the live CaTissue v1.2 system are possible by dropping the database and creating it again, then loading the dump before running 'ant upgrade_db' from the catissue home directory. The csm_migrate_user fix above will need to be re-run each time. However, 'ant upgrade_db' did not successfully upgrade the database. Trying further testing with "upgrade_db_from_suitev1.2_to_plus20" but jboss cache problems are awkward. Connections are held in jboss/server/default/deploy/catissuecore-ds.xml and don't necessarily get updated when a new war file is deployed - which is only a problem if you change the database target between deployments! |
| 31 | |
| 32 | Successful deployment and upgrade required the following process: |
| 33 | Installed db from dump into catissuecore |
| 34 | Checked caTissueInstall.properties |
| 35 | ant deploy_app |
| 36 | ant upgrade_db_from_suitev1.2_to_plus20 |
| 37 | change the column name for the wustlkey problem |
| 38 | check that the LabelGenerator doesn't have the extra definitions which force label generation |
| 39 | move and expand the war file, change the flex configuration class to solve the https / load balancer problem, recompress the war file and copy back to jboss |
| 40 | run the patch as above |
| 41 | test - restart jboss |
| 42 | |