Understanding Snapshot policy
The Snapshot policy feature provides snapshot orchestration at scale (1,024 snaps per storage group). The feature simplifies snapshot management for snapshots.
The flowchart below depicts the basic workflow of using snapshot policies:
Snapshots can be used to recover from data corruption, accidental deletion, or other damage, offering continuous data protection. Many snapshots can be difficult to manage. The Snapshot policy feature provides an end to end solution to create, schedule, and manage snapshots.
The snapshot policy (Recovery Point Objective (RPO)) specifies how often the snapshot should be taken and how many of the snapshots should be retained. The snapshot may also be specified to be secure (these snapshots cannot be terminated by users before their time to live (TTL), derived from the snapshot policy's interval and maximum count, has expired). Up to four policies can be associated with a storage group, and a snapshot policy may be associated with many storage groups. Unisphere provides views and dialogs to view and manage the snapshot policies. Unisphere also calculates and reports on the compliance of each storage group to its snapshot policies.
The following rules apply to snapshot policies:
- The maximum number of snapshot policies that can be created on a storage system is 20. Multiple storage groups can be associated with a snapshot policy.
- A maximum of four snapshot policies can be associated with an individual storage group.
- A storage group or device can have a maximum of 256 manual snapshots.
- A storage group or device can have a maximum of 4000 snapshots.
- When there are 4000 snapshots in existence and another snapshot is taken, the oldest unused snapshot that is associated with the snapshot policy is removed.
- When devices are added to a snapshot policy storage group, snapshot policies that apply to the storage group are applied to the added devices.
- When devices are removed from a snapshot policy storage group, snapshot policies that apply to the storage group are no longer applied to the removed devices.
- If overlapping snapshot policies are applied to storage groups, they run and take snapshots independently.
Unisphere provides compliance information for each snapshot policy that is directly associated with a storage group. Snapshot policy compliance is measured against the count and intervals of the existing snapshots. Snapshots must be valid (must still exist, must be in a non-failed state, and must be at the expected scheduled time). A snapshot could be missing due to it being manually terminated or due to a failure in the snapshot operation.
Snapshot compliance for a storage group is taken as the lowest compliance value for any of the snapshot policies that are directly associated with the storage group.
Compliance for a snapshot policy that is associated with a storage group is based on the number of valid snapshots within the retention count. The retention count is translated to a retention period for compliance calculation. The retention period is the snapshot interval multiplied by the snapshot maximum count. For example, a one hour interval with a 30 snapshot count means a 30-hour retention period.
The compliance threshold value for green to yellow is stored in the snapshot policy definition. Once the number of valid snapshots falls below this value, compliance turns yellow.
The compliance threshold value for yellow to red is stored in the snapshot policy definition. Once the number of valid snapshots falls below this value, compliance turns red.
In addition to performance level compliance, snapshot compliance is also calculated by polling the storage system once an hour for SnapVX related information for storage groups that have snapshot policies that are associated with them. The returned snapshot information is summarized into the required information for the database compliance entries.
When the maximum count of snapshots for a snapshot policy is changed, this changes the compliance for the storage group or service level combination. Compliance values are updated accordingly.
If a compliance calculation is performed during the creation of a snapshot, an establish-in-progress state may be detected. This is acceptable for the most recent snapshot but is considered failed for any older snapshot.
When a storage group and service level have only recently been associated and the full maximum count of snapshots has not yet been reached, Unisphere scales the calculation to the number of snapshots that are available and represents compliance accordingly until the full maximum count of snapshots has been reached. If a snapshot failed to be taken for a reason (such as the storage group or service level was suspended or a snapshot was manually terminated before the maximum snapshot count was reached), the compliance is reported as out of compliance appropriately.
When the service level interval is changed, the compliance window changes and the number of snapshots may not exist for correct compliance.
If a service level is suspended or a storage group or service level combination is suspended, snapshots are not created. Older snapshots fall outside the compliance window, and the maximum count of required snapshots is not found.
Manual termination of snapshots inside the compliance window results in the storage group or service level combination falling out of compliance.
Configuration of alerts related to snapshot policies is available from Settings > Alerts > Alert Policies on the Unisphere user interface.
NOTE: Snapshot policy offsets (the execution time within the RPO interval) and snapshot timestamps are both mapped to be relative to the clock (including time zone) of the local management host. If times are not synchronized across hosts, these appear different to users on those hosts. Even if they are synchronized, rounding that occurs during time conversion may result in the times being slightly different.
|
Unisphere supports the following snapshot policy management tasks:
- Create snapshot policies.
- View and modify snapshot policies.
- Associate a snapshot policy and a storage group with each other.
- Disassociate a snapshot policy and a storage group from each other.
- View snapshot policy compliance.
- Suspend or resume snapshot policies.
- Suspend or resume snapshot policies that are associated with one, more than one, or all storage groups.
- Set a snapshot policy snapshot to be persistent.
- Bulk terminate snapshots (not specific to snapshots associated with a snapshot policy).
- Delete snapshot policies.