-1018 JET_errReadVerifyFailure / Checksum error on a database page
Cause
The DB is corrupted due to a hardware failure.
Resolution
- Evaluate the disk stack, including the motherboard/controller, firmware, connecting cables, and physical drives, and contact the relevant vendors for known issues. Compare your current configuration against the vendors' reference configurations.
- Evaluate whether the problem can be resolved by the latest firmware updates or was triggered by a recent firmware update.
- If some DCs are logging -1018s while other DCs in the same environment aren't, look for differences in hardware configurations.
- Databases that log this error can't be recovered or repaired by integrity checks or semantic database analysis in NTDSUTIL or ESENTUTL.
- Offline defrags may fix the problem in the unlikely case that the problem is due to an index consistency problem.
- Try an offline defrag. Or, restore a system state backup that predates the corruption. Or force-demote, perform a full metadata cleanup, and repromote. If the -1018 error appears, repeat until the hardware root cause is resolved.
When Jet error -1018s occurs on virtualized DCs that are running on the same virtual host only on computers that are using an on-board raid controller, the error can occur because the uninterruptible power supply (UPS) lacked sufficient power for on-board raid controllers to commit changes to disk following loss of electrical power. The workaround is to configure UPS software to shut down virtualized guests upon loss of electrical power. Servers that have dedicated (not on-board) raid controllers with their own battery backups don't experience the -1018 JET error.
