I'm getting DB_PAGE_NOTFOUND on one of my databases, trying to get a
cursor or "stat" on the database. It's still accepting get/put
commands, however.

ie:
DB_PAGE_NOTFOUND: Requested page not found while executing
$dbid stat

I'm assuming this means the database is corrupted? I'm not seeing much
mention of this in the docs, other than as a flag to an environment,
and I'm not using an environment.

-john

If you're not using an environment, this problem could also be caused
by using multiple DB handles -- if you do gets/puts on one handle, and
the stat call on another, they won't see a coherent view of the
database pages (that's what an environment cache provides).

Otherwise, you can find out whether the database file is corrupted by
running the db_verify utility -- does it report any errors?

Regards,
Michael.

db_verify is reporting a corrupted database:
asin.bdb: DB_VERIFY_BAD: Database verification failed

However, I don't see how to fix a corrupted database.

I've read the docs on db_recover, and don't see how it can be used to
repair an individual database file
(http://www.sleepycat.com/docs/utility/db_recover.html).

I must be misunderstanding the docs, or perhaps there's another way to
fix a corrupted database.

FYI, the corrupted database came about from multithreading in Tcl with
the Berkeley DB tcl api. I know that's bad, but I'm also interested in
seeing what I can do when/if a corrupted database occurs in production
(I'm still in development)

-john

db_dump -r [or -R]
delete/rename the db, then do db_load