Appendix A:
A.1.1 Errors Generated by nvmeddbcheck LogError
Media Manager '%s' table failed raw database check (%ld)
This is a serious error; the checker was unable to even check the database tables for simple integrity before moving on to more detailed check. Unless there is some overriding condition (for example, no disk space or memory on the machine), the most likely solution is to recover the last known good NVDB backup and scan in additional backups.
LogInformation
Bad segment found
A missing segment has been detected; this could be achieved by blanking an individual piece of media from a backup. Database compaction will remove the condition but will not recover the missing segment. Incomplete scanning of the media set associated with a backup stream could also be the cause, in which case all relevant pieces of media need to be scanned in.
LogBackground
duplicate xxx entry (Id = %ld, Last Id = %ld) entry out of order (Id = %ld, Last Id = %ld) entry cannot be found (Id = %ld, Last Id = %ld)
These entries indicate an internal discrepancy within the NVDB database; as such they do not cause a problem in themselves but indicate that an
unknown action occurred which compromised the integrity of the database. Whether that relates to scanning in backups, deleting schedules or backups/ media is unknown.
This error indicates that an incomplete database update action occurred because of lack of disk space memory or another reason. It is recommended that the backup logs be reviewed since the last NVDB backup was taken to look for any anomalies. Running a database compaction may help or recovering an previous good NVDB backup may be required.
LogInformation
Media Mid from two generations (Id = %ld, Last Id = %ld)
Can be caused by the re-use of media, possibly intertwined with an NVDB recovery; this is a benign problem.
LogError
Internal error walking database table
This is a serious error; the checker was unable to even check the database tables for simple integrity before moving on to more detailed check. Unless there is some overriding condition - for example, no disk space or memory on the machine, the most likely solution is to recover the last known good NVDB backup and scan in additional backups.
LogWarning
Backup %ld stream %s is incomplete
Session with incomplete new streams Backup %ld Job %ld Instance %ld Phase %ld
Backup %d is incomplete. Required old streams are missing
The information that relates to the backup does not contain all the required information. Run a database compaction to remove the backup from the database and then recover the backup, if required. This is a benign problem. LogInformation
No Session matching Backup Index %ld Backup Index %ld is for a deleted session
There is a backup index in the database for which there is no job session record; possibly the index has been scanned in from tape or the jobs have all been deleted. Run a database compaction to remove the backup from the database and then recover the backup, if required. This is a benign problem. LogInformation
Backup %ld has corrupt index file
The index file is not readable; it needs to be deleted by hand (the path to the file is now indicated in the log). The index can be scanned in again from the backup device.
LogBackground
The saveset index file is offline or missing for %ld backups
The index can be scanned in again from relevant backup device. LogInformation
Saveset index files found for %ld deleted backups
Index files have been found for unrecorded backup - maybe some backups have been scanned in to recover the index files. Run a database compaction. LogWarning
Failed to check..
There has been a problem in the checking of the database tables; possibly due to lack of resources. This needs to investigated because no checking of the content is able to occur unless basic access is possible. This is similar to a problem with checking the raw integrity of the database. However, it does not necessarily indicate a problem with the database(s).
LogWarning
Media Manager database check identified unreferenced records
Some problems were identified. Run a database compaction. LogWarning
Failed to compact of media manager database '%s' records Failed to compact media manager database
Ordinarily a compaction should run successfully, run nvmeddbcheck manually to identify the problem area.
LogWarning
Warning 2008/02/26 05:25:00 1322 Media NetvaultServer Backup 50257 has corrupt index file
Re-scan indexes from the backup device or recover a good NVDB, and re- scan the additional backups from tape.
LogBackground
Check Media table Mid index
Media: Mid from two generations (Id = 57328, Last Id = 57224)
Can be caused by the re-use of media, possibly intertwined with an NVDB recovery; this is a benign problem. After being marked for re-use, the MID generations will be incremented and when re-used again, incremented again. Normally, the records related to the use of MID would be cleared but there's information in the database that relates to the same MID from different times. Possibly recovering an NVDB or scanning in a duplicated (older) copy of the tape would cause this. It is not a problem.