Bug #4430
closedhsql database name and port in Kepler should be configurable
0%
Description
I can't start Kepler GUI when I have hadoop server running. It complains:
Caused by: java.sql.SQLException: Failed to connect to url "jdbc:hsqldb:hsql://localhost/hsqldb" as user "sa. Perhaps there was an error launching the db server. In addition, sometimes the database cache can become corrupt, so another solution is to remove the "/Users/jianwu/.kepler/cache/cachedata/hsqldb" directory.
[run] at org.kepler.util.DBConnectionFactory.getDBConnection(DBConnectionFactory.java:126)
With Dan's help, I found out it is because hadoop also have its hsql databases. Now I work it around by setting port of hsql server to another one (e.g. 9002)
1) add 'server.setPort(9002);' at launchDBServer() function.
2) change url to url = "jdbc:hsqldb:hsql://localhost:9002/" + dbName;
So I think hsql database name and port in Kepler should be configurable.
Other possible changes could be better exception handling, such as automatic port change, or mentioning this port is already in use.
Updated by Chad Berkley about 15 years ago
Not critical for users so moving to post 2.0.
Updated by Daniel Crawl almost 14 years ago
The ports and names are set in the configuration files:
common/resources/configurations/configuration.xml
core/resources/configurations/configuration.xml
provenance/resources/configurations/configuration.xml