Backup Troubleshooting - SQL Server iDataAgent
The following section provides information on troubleshooting backups.
|Database name contains [ and ] brackets||The use of embedded brackets '[' and ']' in database name may result in backup failures.|
|All data paths for the subclient are offline or busy|| This error may be displayed if the Override Datapaths option is selected in the Data Paths dialog box in the Subclient Properties for a Log Storage Policy. This results in the Transaction Log backup operation waits for resources.
To work around this issue, deselect the Override Datapaths option.
|Time Out Failures||The default time allocated for backup and restore operations of SQL databases is 0 (infinite). If a backup or restore operation fails due to a timeout being reached, you can configure the nSqlQueryTimeout registry key to increase the amount of allocated time for backup or restore operations.|
|SQL Server jobs that cause backups to terminate|| There are a few jobs that SQL Server restricts during a backup. If one of these jobs are initiated while a backup is already in progress (or if a backup is initiated while one of these jobs is in progress) the backup job will terminate. These jobs are:
|Backup chain is broken|| When a full or differential backup is performed outside of the system, for example, from SQL Enterprise Manager, the subsequent log backups performed using SQL Server iDataAgent is set Do not convert log backups to full if log backup was performed using other software in the Subclient - Backup Rules tab.
Make sure to enable the Disable Log Consistency Check in the Subclient - SQL Settings tab to ensure that the backup job completes successfully.
Completed with one or more errors
Backup jobs from Microsoft SQL Server iDataAgent will be displayed as "Completed w/ one or more errors" in the Job History in the following cases:
- When a subclient which contains multiple databases is backed up, if one of the database gets deleted from the sql server, then that database is not backed up and the remaining databases get backed up.
- For a default subclient, if all the databases that are its content are auto discovered, then even if one of the database has been deleted from the sql server, the job completes successfully as the database that has been deleted from the sql server will be removed from the default subclient content. But if the database that has been deleted from sql server is part of the default subclient, then the database is not removed from the subclient content.
- When a subclient which contains multiple databases is backed up, if one of the database is not backed up due to reasons like, database is in standby mode or database got corrupt etc, then job completes w/ one or more errors. The databases that failed will be shown as part of the failed items and those that were backed up will be shown as part of the successful items.
- When running a backup, a check is made to verify if the backup is restorable. If the log chain is broken (e.g., when a log backup is ran outside of the software) or if there are no full backups for a corresponding differential backup then the backup of the database fails and the job will complete with errors. A Job Pending Reason (JPR) explains why the backup failed. In the next backup attempt, this database will be backed up as a full database. An alert can also be configured for this job.
- A SQL backup job for a subclient with multiple databases will not retry backing up a single database if it fails. However, the job status will be displayed as Completed With Errors.
If the job goes into pending state, the job will restart from the point where it failed and if an attempt to back up the failed database has already been made, another attempt will not be performed.
- For databases that are manually defined in a subclient but are inaccessible (e.g., it is not recognized, has been deleted, etc.), the job status for the backup will be displayed as Completed With Errors.
An event will be created for the inaccessible database during backup. If the inaccessible database is not needed, it can be permanently deleted from the subclient content.
- When a backup is run, a check is made to verify if the backup is restorable. If the log chain is broken (e.g., when a log backup is run outside of the software) or if there are no full backups for a corresponding differential backup then the backup of the database fails and the job will be displayed as Complete With Errors.
A Job Pending Reason (JPR) explains why the backup failed. In the next backup attempt, this database will be backed up as a full database. An alert can also be configured for this job.