From: Blackwell <bla...@ne...> - 2008-04-30 07:12:43
|
Jesse Pavel wrote: > Hi Shlomy, > > If the problem is related to the backup replication of the > large DB, you could try building a custom HSQLDB by first > changing the backupFile() and renameBackupFile() methods to > no-ops in DataFileCache.java (starting at line 638) in the > HSQLDB source. It seems silly that they didn't include a user > option, so far as I can see, to avoid ZIP'ing hundreds of > megabytes on every shutdown. > > I'm not sure that would solve the problem, or if you'd want > to get that involved in the DB, but it may be an idea. Jesse, being a solution that can be tested in what seems like minutes, I'd definitely take this route, if only as a first step for analysis or further changes towards another db engine or db interface. I'd probably try to find the spot for the HSQLDB JDBC URL parsing and add a custom option there. (All that assuming that I am not missing something about this seemingly odd mechanism in HSQLDB. I have a gut feeling that this cannot really be how the creators meant it to behave, but oh well.) Bye bye |