[
https://issues.apache.org/jira/browse/DERBY-3541?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12579332#action_12579332 ]
Jørgen Løland commented on DERBY-3541:
--------------------------------------
I guess this (quiting derby without shutting down the databases) has other implications as well. E.g., when a booted database is not properly shut down, the checkpoint that is normally made on shutdown will not be there, increasing the time to recover the next time that database is booted.
Combining what you say here with the discussion on 3447, my suggestion is twofold:
* I agree that derby should not hang if the user shuts down an application without shutting down the databases. Your suggestion of using daemon threads for replication sounds good to me.
* On the other hand, we control the ij application, and we may want to improve it. Documenting this problematic behavior like suggested may be enough, but it would also be possible to abort the 'quit;' call with an exception saying:
ij > quit;
Error: could not quit ij because the following databases have not been shut down: "test1", "test2". Shut down these databases first and then quit ij.
I don't feel strongly about either option (only document / fail 'quit;'), but I think that in any case, the severity can be set lower now that we know the cause of the problem.
> quit; on ij with authentication enabled does not shutdown the derby modules that are loaded
> -------------------------------------------------------------------------------------------
>
> Key: DERBY-3541
> URL:
https://issues.apache.org/jira/browse/DERBY-3541> Project: Derby
> Issue Type: Bug
> Components: Security
> Affects Versions: 10.4.0.0
> Reporter: V.Narayanan
> Attachments: derby.properties, PrintStackTrace.diff, WithoutAuthentication_StackTrace.txt
>
>
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.