db-derby-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Knut Anders Hatlen <Knut.Hat...@Sun.COM>
Subject Re: [Proposal for Review] The fix for DERBY-3347 is important enough for the Development Community to issue a recommendation the User Community to upgrade.
Date Tue, 29 Apr 2008 13:56:06 GMT
Andrew McIntyre <mcintyre.a@gmail.com> writes:

> Funny, I'm having trouble finding the exact mail that convinced me
> that we should issue a release for the issue, but a thorough reread of
> DERBY-3347 should do it for you. Expecting users to read DERBY-3347
> carefully and understand what is going on is not something we should
> expect them to do. :-)
>
> So, we should provide a more descriptive summary of the issue. What
> was happening, how to figure out if you've hit it, possible remedies,
> what was done to fix it, etc. Our release note template already
> provides a good template for the information that a user would want to
> see.
>
> Does anyone have time to write a good release note for DERBY-3347?

I have just attached a release note to DERBY-3347. It mentions the
affected versions, some of the possible symptoms, and it points to the
wiki page that describes the consistency checking. Anyone should feel
free to improve it and make it a "good release note".

> It would be nice to have a documented way for a user to identify if
> they've hit this problem, whether it's on a wiki page, in the release
> notes, or in the announcement itself. The wiki page has a good
> description of how to run the consistency check and an example of
> output. Maybe the wiki page could be augmented with the output that
> might be expected if a user has hit this issue with one of their
> databases.

I don't think there is any method to say for sure that you've hit this
particular problem. The consistency check will give an error message,
but the exact message depends on which part of the database was
corrupted and how. My guess is that these two errors are the most likely
ones to be seen during consistency checking:

ERROR XSDB3: Container information cannot change once written: was 0, now 80
ERROR XSDG2: Invalid checksum on Page Page(0,Container(0, 1313))

But these messages could also indicate that corruption has occured
because of some other bugs (for instance DERBY-3606/DERBY-3707) or disk
failures.

I've never been able to reproduce the corruption issue, though, so I
can't say for sure which message you would get.

-- 
Knut Anders

Mime
View raw message