Table of ContentsView in Frames

What can happen with a fractional space-reserved volume


When a LUN is fully space reserved, write operations to that LUN are guaranteed against failure caused by an out-of-space condition due to Snapshot copy disk space consumption. When the overwrite reserve for a volume is set to less than 100 percent, however, write operations to the LUNs on that volume are no longer guaranteed when the storage system volume runs low in free disk space due to Snapshot copy space consumption.

Attention: If a storage system volume runs out of overwrite reserve space, write operations to a LUN on that volume will fail and the host application may terminate, report I/O errors, or exhibit unexpected behavior.

Data ONTAP provides two space management tools to ensure that a fractionally space-reserved volume does not run out of overwrite reserve: automatic FlexVol volume expansion and automatic Snapshot copy deletion from FlexVol volumes. These features, summarized in the following paragraphs, monitor the reserved space and take action if the free space becomes scarce. For more detailed information, see the Data ONTAP SAN Administration Guide for 7-Mode.

Automatic expansion of flexible volumes Data ONTAP can automatically expand the volume that is used to store Snapshot copy data, provided the volume is a flexible volume with the guarantee option set to volume. When the flexible volume is nearly full, Data ONTAP automatically expands the volume into the space preallocated for it in the aggregate. The automatic Snapshot copy deletion and FlexVol volume expansion features can be enabled separately, or together with one policy to be applied before the other. When fractional-space-reserved volumes hold LUNs that store SQL Server database files, however, only the automatic FlexVol volume expansion feature can be used, if needed.

Automatic deletion of Snapshot copies from flexible volumes Data ONTAP can automatically delete one or more Snapshot copies on the volume, provided the Data ONTAP Snapshot copy autodeletion policy is enabled and set to trigger when the overwrite reserve is nearly full on the volume. If the trigger condition is detected, the oldest or newest Snapshot copies are deleted until a configured percentage of the volume is free space. If you do not want to automatically delete SnapShot copies on the volume, you can set the overwrite reserve to 100 percent, by setting the fractional space reserve to 100 percent on the storage system. Note that this Data ONTAP feature is not designed specifically to support backup and restore operations on SQL Server databases:

Attention: Because SnapManager is not aware of Snapshot_Autodelete, autodelete might delete all SnapManager backups including the most recent backup. This should be taken into consideration when deploying Snapshot autodelete.

SnapManager is not aware of the Snapshot_Autodelete process which can be defined by the storage administrator on the NetApp controller. If an autodelete occurs in Data ONTAP and a Snapshot copy is deleted that is part of a SnapManager backup, SnapManager will detect that the original SnapManager backup is invalid. SnapManager will not show this backup in the GUI and you will not be able to restore that backup. The metadata will get deleted when you run backup with the retention policy or during other backup delete operations. During the backup deletion, if SnapManager for SQL Server finds those metadata are useless (in other words, they are not associated with a remote backup), then they are deleted at that time.