[Please CC any replies, I'm subscribed nomail]
As promised I've given it a bit of polish and it's actually almost useful.
You can have a look at it http://svana.org/kleptog/pgsql/pgfsck.html
Just unpack the files into a directory. It's just a perl script with two
modules so no compiling necessary. You can download the package directly at
http://svana.org/kleptog/pgsql/pgfsck-0.01.tar.gz
I've tested it on versions 6.5, 7.0 and 7.2 and it works. It shouldn't
crash, no matter how bad a file you feed it. It can output insert statements
also to help reconstruction.
Here is an example of the program being run over a suitably hexedited file.
# ./pgfsck -r 16559 kleptog website
-- Detected database format 7.2
-- Table pg_class(1259):Page 1:Tuple 0: Unknown type _aclitem (1034)
-- Table pg_class(1259):Page 1:Tuple 49: Unknown type _aclitem (1034)
-- Table website(16559):Page 0:Tuple 7: Tuple incorrect length (parsed data=57,length=1638)
-- Table website(16559):Page 0:Tuple 44: Decoding tuple runs off end: 627338916 > 69
-- Table website(16559):Page 0:Tuple 70: Bad tuple offset. Should be: 3784 <= 11592 < 8192
Currently the following features are not supported:
- Toasted / compressed tuples
- Checking indexes doesn't work (should it?)
- Views just produce empty output (because they are)
- Arrays don't work
- Since each type output has to be written, many types are not correctly output
- Split tables (1GB) are not supported past the first part.
- Some system tables in some versions have a strange layout. You may get many harmless warnings about the formats of
pg_class,pg_attribute and/or pg_type.
Most of these are basically because I don't know how they work, but with a
bit of work some of these should be fixable.
Have a nice day,
--
Martijn van Oosterhout <kleptog@svana.org> http://svana.org/kleptog/
> Canada, Mexico, and Australia form the Axis of Nations That
> Are Actually Quite Nice But Secretly Have Nasty Thoughts About America
