|Message:||%1 (%2) %3The database engine lost one page of bad data. It is highly recommended that an application-level integrity check of the database be run to ensure application-level data integrity.|
This Warning event indicates that a database repair process may have caused data to be lost. Although the repair process tries to validate all system tables and indexes, any data that cannot be repaired is discarded. If the database cannot be repaired, an error occurs.
When a database is in an inconsistent condition and a repair is performed by using eseutil /p, ESE event 500 may be logged during the repair operation. ESE event 500 indicates that the repair process found a bad page and tried to repair it, but the repair of the page was unsuccessful and the page was discarded. ESE event 500 is a serious error that frequently indicates that the database is damaged.
To resolve this warning, run isinteg -fix on the database. After a successful repair that indicates some lost data, run eseutil /d to defragment the database and rebuild indexes, and then run isinteg -fix to restore referential integrity between tables in the database. Repair may not have corrected all problems in the database. To completely guarantee database integrity, you may want to use the Move Mailbox tool to move all mailboxes temporarily to a different database. You can then delete the files for this database. Exchange will create new database files when you mount the database, and you can then move mailboxes back to the database. If this is a public folder database, replicate all folders to a different public folder database and verify that replication has completed for all folders. You can then delete the files for this database. Exchange will create new database files when you mount the database, and all folders will automatically replicate back to this database after a period of time.
If the repair process fails because the database is too damaged even for a repair, you must restore from Exchange-aware online backup. If there is no valid backup available, check to determine if users are storing their data locally in .pst files on their client computers. If they are, create a new blank database and use the .pst files to recover as much data as possible.