Backup of a file server may fail with error Phoenix 322: Database disk image is malformed.
You can run into this issue in the following scenarios:
- Phoenix folders and processes are not excluded from the antivirus scans.
- The local state map database residing on the server is corrupt.
- Follow the instructions in the Logs article and obtain the job logs.
- Extract the logs. Access the Phoenix<YYYYMMDD>_<timestamp>.log to determine if the following excerpt is present in the log.
level=error ts=2020-02-22T23:27:48.3044221-07:00 stack="database disk image is malformed\\ngoroutine 247 [running]:\nruntime/debug.Stack(0xc00db8cfa0, 0xc00010a380, 0x20)\n\tC:/Go/src/runtime/debug/stack.go:24 +0xa4\nphoenix/pkg/client/logutil.Error(0xd34040, 0xc00db8cfa0, 0xc5a0b7, 0x18, 0xc00db33d00, 0x4, 0x4)\n\tF:/gitlab/builds/8f2101d3/0/phoenixdev/phoenix/go_workspace/src/phoenix/pkg/client/logutil/logger.go:76 +0x25e\nphoenix/fileserver/client/upload.(*Uploader).smapCommitWorker(0xc0001f5340, 0xd48dc0, 0xc0002e8130, 0xc00034b8c0)\n\tF:/gitlab/builds/8f2101d3/0/phoenixdev/phoenix/go_workspace/src/phoenix/fileserver/client/upload/smap_worker.go:63 +0x66b\ncreated by phoenix/fileserver/client/upload.(*Uploader).startSmapCommitWorkers\n\tF:/gitlab/builds/8f2101d3/0/phoenixdev/phoenix/go_workspace/src/phoenix/fileserver/client/upload/smap_worker.go:14 +0xf1\n"
- Exclude the Phoenix folders and processes from the antivirus scan
Most antivirus programs lock frequently accessed files before scanning them. Preventing any writes to database files while the antivirus scans them can cause database corruption and lead to this database error. Follow the instructions in this article to whitelist Phoenix.
- Check for NTFS disk errors in the event logs
Contact your hardware vendor to correct issues with the disk if you see NTFS disk-related errors in the event logs. Damaged or corrupt system files can also cause this issue. Sudden system shutdown due to power failure or a forced restart can result in corrupt files and make the system hard disk inaccessible.
- If the issue persists, perform the following tasks:
Stop the Druva Phoenix Agent Client service on the server.
Navigate to C:\ProgramData\Phoenix\FS\state\backup\<backupsetID>\
You may see a C.db file. This is the local database file.
Rename it to C.db.old
Trigger the backup operation again.
Note: Renaming the state map database will trigger a full scan during the next backup to build a new database. If you have implemented a backup window according to the time your average incremental backup takes to complete, your job may fail with “Backup Window Expired.” Running a manual “backup now” will allow the backup window to be ignored and let the full scan complete allowing the backup job to run to completion.