Subject | Re: [firebird-support] Corrupt database |
---|---|
Author | Ann W. Harrison |
Post date | 2007-10-26T14:31:33Z |
Fabiano Bonin wrote:
of a crash at a bad moment. When a new page is allocated, it
is first removed from the free page list, then added to whatever
structure it will become. If there's a crash after the PIP
(free space list) page is written but before the page is added
to a table, index, or whatever, it becomes an orphan. No
information is lost, but that space will be unavailable until
the database is backed up and restored. The same thing can
happen when a page is released - it's removed from its table
(index or whatever) first, then added to the free list. A
crash between those steps also orphans pages.
when the database is restarted. It's an interesting and serious
error, but not a corruption.
Best regards,
Ann
>The error "Page <nnn> is an orphan" can be a benign artifact
> server1.romanatecidos.com.br Fri Oct 26 10:00:12 2007
> Database: //romana.fdb
> Page 1418989 is an orphan
of a crash at a bad moment. When a new page is allocated, it
is first removed from the free page list, then added to whatever
structure it will become. If there's a crash after the PIP
(free space list) page is written but before the page is added
to a table, index, or whatever, it becomes an orphan. No
information is lost, but that space will be unavailable until
the database is backed up and restored. The same thing can
happen when a page is released - it's removed from its table
(index or whatever) first, then added to the free list. A
crash between those steps also orphans pages.
>That should be a transitory runtime error - and should not reoccur
>
> server1.romanatecidos.com.br Fri Oct 26 10:00:12 2007
> Database: //romana.fdb
> internal gds software consistency check (page in use during
> flush (210), file: cch.cpp line: 3653)
>
when the database is restarted. It's an interesting and serious
error, but not a corruption.
Best regards,
Ann