Resolving Snapshot Difference problems

To use the Snapshot Difference feature, you must first set up a NetApp user ID and password on the client. The user ID and password are necessary for IBM® Storage Protect to connect to the filer. Set up a user ID/password with root authority on AIX® and Linux®, or one with administrative authority on Windows. Set the authority level to be the same as the authority level used when you map or mount the filer volume. Ensure that you use the fully qualified host name or the dotted IP address format for the filer name. Issue the backup-archive client SET PASSWORD command to save this user ID/password information.

The Snapshot Difference feature supports the following features, which are only applicable at the volume level:

• NetApp/N-Series filers that are running Data ONTAP release 7.3 to 9.9.1
• Common internet files system-attached (CIFS) volumes
• Both traditional and FlexVol filer volumes
• Java™ GUI
• Network file system (NFS) attached volumes

The Snapshot Difference feature does not support the following features:

• SAN-attached NetApp/N-Series volumes
• QTrees or subdirectories
• VFiler volumes with a filer that is running ONTAP V8.1.0 or earlier are not supported. Vfiler volumes with a filer that is running ONTAP V8.1.1 or later are supported.

The lack of Unicode support from NetApp prevents IBM Storage Protect from processing any files that use characters that are not in the 7-bit ASCII range. IBM Storage Protect can back up only names that contain ASCII characters. Two Snapshot Difference behaviors were noted when testing with Unicode characters:

1. Snapshot Difference incremental command ends with return code 13001. This return code happens with the 'specials' and 'surrogate' ranges of Unicode for Snapshot Difference filer volumes that are created with the UTF8 flag. This Snapshot Difference error happens more frequently without the UTF8 flag. IBM Storage Protect ends with error message ANS5283E The operation was unsuccessful. No files are backed up.
2. Snapshot Difference application programming interface (API) does not fail, but returns characters that are not part of the real name. IBM Storage Protect inspects the string to see whether any character is outside of the 7-bit ASCII range. If so, IBM Storage Protect skips the file and logs the error to the dsmerror.log file.

• You exclude a file by adding an exclude rule in the include/exclude file. IBM Storage Protect performs a backup of the current snapshot with that exclude rule in effect. You did not change the file, but do remove the rule that excluded the file. A snapshot-assisted incremental backup command with the snapdiff option does not detect this incl/excl change because it detects file changes only between two snapshots. The files themselves must be changed in order for the Snapshot Difference API to detect the change and for IBM Storage Protect to back up the file.
• You added an include statement to the options file. This include statement takes effect only if the file is detected to have changed by the Snapshot Difference API. The files might not get backed up because IBM Storage Protect is not inspecting each file on the volume during the backup operation.
• You explicitly delete a file from IBM Storage Protect inventory by issuing the DSMC DELETE BACKUP command. The Snapshot Difference API does not detect that a file was manually deleted from IBM Storage Protect by you. Therefore, the file remains unprotected in storage. The file is unprotected until it is changed on the volume and the change is detected by the Snapshot Difference API. After the change is detected, the Snapshot Difference API signals IBM Storage Protect to back up the file again.
• Policy changes such as changing the policy from Mode=modified to mode=absolute are not detected. The entire file space is deleted from inventory. The undetected policies cause IBM Storage Protect to create a snapshot to use as the source (base) and a full incremental backup is performed.

Running a full incremental backup without the snapdiff option solves these limitations. IBM Storage Protect does not control what constitutes a changed object. The changing of objects is now controlled by the Snapshot Difference API. Therefore, running a full incremental backup without the SNAPDIFF option ensures that all file changes are detected.

You can use the following trace flags for Snapshot Difference processing:

• enter
• exit
• general
• snapshot
• hci
• hci_detail
• diskmap
• diskmap_detail
• hdw
• hdw_detail
• bacache
• snapdiffdb