-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Our 2.7.1-based production server has started throwing up a lot of POSKeyErrors recently. Forcing regeneration of the Data.fs.index didn't fix it. Packing didn't fix it or generate any new errors - we haven't packed the database since it went live. fsrecover.py doesn't do anything. Running fsrefs.py on the Data.fs gives *scary* long output (attached both unpacked and packed ouput). I haven't been able to get checkbtrees.py to produce any useful output (it doesn't appear to want to poke into the Application). I'm not sure it'll tell me any more than fsrefs.py though. - From my reading of the fsrefs.py output, it looks like the corruption is the result of a single transaction. I'm more than happy to look into this further, and also to supply the Data.fs (200Mb unpacked, 20Mb packed) to help analyse it. Richard -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.4 (GNU/Linux) iD8DBQFBEDKUrGisBEHG6TARAo4cAJ9xuXtYf95uJf/oj+95CfdFJ9qNTgCfaRT7 xETqjHTNProIG6n1Yr1vjp8= =moPK -----END PGP SIGNATURE-----