A duplicate image of a volume at a certain point in time is referred to as a snapshot. If you unintentionally update copy-source volume data, you can restore the data from a previously taken snapshot. The copy-source volume is referred to as the primary volume (P-VOL), and the copy-destination volume is referred to as the secondary volume (S-VOL). When the P-VOL is updated, only the data existing before updating in the update-target area is copied to the storage pool as data in snapshot volumes.
By storing only difference data between P-VOL and S-VOL, this scheme reduces capacity and allows a cost-efficient copy to be created. Furthermore, because data is shared between S-VOLs, less capacity is used even if multiple generations of snapshots are taken.
Because the storage pool is shared, there is no need to create a dedicated storage pool for snapshots.
-
You should back up any necessary data to another device (for example, via the host) with the following in mind. Also, monitor the used capacity of the storage pool to secure sufficient free space.
-
The S-VOL uses the data in the P-VOL. Therefore, if a failure occurs in the P-VOL and the data of the P-VOL becomes unavailable, the data of the S-VOL also becomes unavailable.
-
Difference data is stored in the storage pool. If the storage pool becomes unavailable due to a failure or the storage pool becoming full, data in snapshot volumes will be lost.
-
If data in snapshot volumes is lost during restore (that is, during execution of an operation to restore P-VOL data from S-VOL data), data in the restore-destination volume (P-VOL) also becomes unavailable.
-
You can take snapshots of a volume with data reduction enabled (P-VOL) to create snapshot volumes (S-VOLs). However, the data reduction function is not enabled for the S-VOLs. Consider that when managing storage pool capacity.
-
If event log KARS06170-C is issued, the snapshot volume status related to the target volume (P-VOL) status changes to "IOSuppressed" and I/O operations to the volume are suppressed. Delete all the snapshot volumes created from the target volume (P-VOL). Then, delete the target volume (normal volume at this point). I/O operations are not possible even when the target volume becomes normal. (The volume status remains "IOSuppressed.")
-
-
A snapshot cannot be restored by restoring the storage cluster configuration according to the procedure in Performing the configuration backup operation (Bare metal). A system administrator or users need to delete all the restored snapshot volumes and volume paths.
-
Snapshot volumes cannot be expanded. To expand volumes whose volumeType is "Snapshot" and whose snapshotAttribute is "P-VOL", delete all the snapshot volumes created from the applicable volumes. Volumes whose volumeType is "Snapshot" and whose snapshotAttribute is "P/S-VOL" or "S-VOL" cannot be expanded.
A snapshot configuration consisting of the copy source and copy destinations as follows is referred to as a snapshot tree. A different generation of the S-VOL is created by the time at which a snapshot is taken.
Another snapshot can be created by using an already created snapshot as the copy source. Another snapshot can be created by using an already created snapshot as the copy source. This snapshot configuration is called a cascade configuration. A cascade configuration can comprise of a maximum of 64 layers. In a cascade configuration, a volume can have both the P-VOL and S-VOL attributes. Such a volume is called a P/S-VOL.
A cascade configuration with a maximum of 64 layers can be created under the P-VOL layer. For example, the cascade configuration in the following figure has three layers.
The following table lists the operations on snapshots.
Operation |
Description |
Reference |
---|---|---|
Prepare to take a snapshot |
Create a volume (S-VOL) for taking snapshots. When you prepare to take a snapshot, a volume (S-VOL) will be created. A snapshot cannot be taken without creating a volume (S-VOL). |
Preparing to take a snapshot |
Take a snapshot |
Take a snapshot for the prepared S-VOL. |
Taking a snapshot |
Delete snapshots |
Delete snapshot volumes that are no longer needed. |
Deleting snapshots |
Restore P-VOLs from a snapshot |
Restore P-VOL data from S-VOL data. |
Restoring P-VOLs from a snapshot |
Obtain P-VOL information for the volume |
Obtain information about the P-VOL from which S-VOL was created. |
Obtaining P-VOL information for the volume |
Obtain a list of S-VOL information for the volume |
Obtain a list of S-VOL information created from P-VOL. |
Obtaining a list of S-VOL information for the volume |
Modify settings of individual volumes |
Edit the settings for the volume. |
Modifying settings of individual volumes |
Snapshot specifications
The following table lists the specifications for snapshots.
Item |
Specifications |
---|---|
Number of S-VOLs (per P-VOL) |
1,023 |
Number of volumes (per storage node) |
Total number of P-VOLs, P/S-VOLs, S-VOLs, and ordinary volumes 1,024 |
Number of volumes (per system) |
|
Maximum total volume capacity when snapshots are used (per storage controller) |
Approx. 4 PiB If you prepare to take snapshots of approximately 4 PiB or larger, the operation might fail with KARS06002-E. |
Maximum number of simultaneous operations of volume operations and snapshot operations |
Approx. 20 If you perform more than 20 operations at the same time, operations might not be performed successfully due to a timeout. Volume operations refer to volume creation, deletion, expansion, and setting edition. Snapshot operations refer to preparation to take snapshots, taking, deleting and restoring snapshots. The operation target also includes volumes with data reduction enabled. A single operation handling multiple volumes is counted as one operation. For example, a volume creation operation with 20 or more volumes specified is counted as one operation. Completion of the job means completion of the operation. When the job state is "Succeeded", the operation is complete. |
Snapshot volume name |
Number of characters: 1 to 32 Characters that can be used: Numbers (0 to 9), uppercase alphabet (A to Z), lowercase alphabet (a to z), symbols (- , . : @ _) |
Maximum number of layers in a cascade configuration |
64 |
* (Cloud) In Multi-AZ configuration, the tiebreaker node, which does not have capacity (no drives exist), is not included in the number of storage nodes. |
Snapshot-specific information obtained from volume information
You can obtain the following snapshot-specific information by obtaining a list of snapshot information.
Type of information |
Description |
---|---|
snapshotStatus |
Snapshot status indication: Normal, Empty (empty status without snapshot meta data)*, Deleting, Restoring, Preparing (to take snapshots), Prepared, or Error An ordinary volume is indicated by "-". * This is a temporary status when creating or deleting S-VOLs. When you create S-VOLs, the status transitions to Preparing after a while, and the applicable volumes are deleted when you delete S-VOLs. You can also obtain snapshotStatus from a list of volume information. |
snapshotProgressRate [%] |
Progress rate of snapshot operation. Indicated only when snapshotStatus is Preparing, Restoring or Deleting. In the case of other statuses, null is output. snapshotProgressRate is not displayed at the same time as the match rate (snapshotConcordanceRate [%]). You can also obtain snapshotProgressRate [%] from volume information. |
snapshotTimestamp |
Time when the snapshot is taken. This is displayed when snapshotStatus is "Normal." In the case of other statuses, null is output. You can also obtain snapshotTimestamp from volume information. |
snapshotType |
"Snapshot" is displayed for a snapshot volume. You can also obtain snapshotType from volume information. |
snapshotConcordanceRate [%] |
Match rate [%] between this volume and the comparison-target volume. When this volume is the latest generation S-VOL, the volume to be compared will be the P-VOL. When this volume is not the latest generation S-VOL, the volume to be compared will be an S-VOL of a newer generation. For example, if the P-VOL is written in a multi-generation configuration (1:N configuration for a P-VOL and S-VOL), only the latest generation of S-VOL will have a lower match rate. If the S-VOL is written, the match rate will decrease only for the S-VOL of the written generation. The match rate is an approximate value and might have a large margin of error. Differential data indicated by the match rate is managed in units of 256 KiB. When restoring and deleting snapshots, the larger the amount of data difference between P-VOL and S-VOL is, the longer the process will take. This information can be used to estimate the processing time for snapshot restoration and deletion in advance. You can also obtain snapshotConcordanceRate [%] from volume information. (Bare metal) For details about the approximate processing time for snapshot operations, see Processing time for snapshot operations (Bare metal) in this document. |
isWrittenInSvol |
This information shows whether the host has written to the S-VOL. This is displayed when snapshotStatus is "Normal." In the case of other statuses, null is output. You can also obtain isWrittenInSvol from volume information. |
Status transition (snapshot operation and snapshotStatus)
The snapshot Status transits depending on the snapshot operation.
It might take about one minute for snapshotStatus to transit. To confirm the transition of snapshotStatus, wait about one minute and then check again.
Whether the volume is a snapshot volume can be found in snapshotAttribute of volume information. "P-VOL", "P/S-VOL", or "S-VOL" is indicated. An ordinary volume is indicated by "-".
Also, the number of S-VOLs in the P-VOL or P/S-VOL can be found in numberOfSnapshots in volume information.
Whether the P-VOL and S-VOL can be written or read
The P-VOL is readable/writable regardless of snapshotStatus of the copy-destination volume.
The S-VOL is readable/writable only when snapshotStatus is "Normal."
Maximum pool capacity consumed by snapshots
You can calculate the maximum pool capacity that is consumed by snapshots for each P-VOL by using the following formula. To calculate the total maximum capacity, calculate and sum the capacity for all P-VOLs.
Maximum consumed capacity of a storage pool for each P-VOL [MiB]=(1) × 42 + (2) × 175,434
-
(1): Result of the following calculation rounded up to the nearest integer
(P-VOL capacity [MiB] × number of snapshots × 2 / 42)
-
(2): Result of the following calculation rounded up to the nearest integer
(P-VOL capacity [MiB] × number of snapshots × 2 / 2,921,688)
The number of snapshots in the above formula is not the number of current snapshots but the maximum number of snapshots that have been created.
Notes on storage pool capacity
The storage pool capacity minus the virtual volume management area is the capacity that can store data in that storage pool. The virtual volume management area is approximately 3% of the storage pool capacity.
The capacity of the virtual volume management area varies depending on the capacity of the P-VOL, the number of snapshots, and other factors. The estimated capacity for each P-VOL is calculated using the following formula. The capacity for each P-VOL is calculated using the following formula. To calculate the capacity of all the virtual volume management areas, calculate and sum the capacity of all P-VOLs.
Capacity of the virtual volume management area for each P-VOL [MB] = (4 + (1)) × 42
-
(1): Result of the following calculation rounded up to the nearest integer
( 2 × update type × number of snapshots + 64 ) × (volume capacity [MiB] / 3145728 [MiB] )
-
Update type: Specify 1 or 2 depending on the volume to be updated.
-
Specify 1 to update only the P-VOL.
- Specify 1 to update only the S-VOL.
- Specify 2 to update both the P-VOL and S-VOL.
-
-
Number of snapshots: Number of S-VOLs per P-VOL
-
Volume capacity: Capacity of the P-VOL [MiB]
If the capacity of the storage pool is insufficient, the storage pool might become full and the data in snapshot volumes might be lost. Therefore, when creating a storage pool, it is necessary to estimate the copy capacity (capacity of the differential data of the snapshots to be taken) and set the storage pool capacity to be sufficient to accommodate the copy capacity. The capacity of the storage pool is always consumed in units of 42 [MiB]. Therefore, even if the capacity of the difference data is less than 42 [MiB], the storage pool will still need 42 [MiB] of space.
Pay attention to the free capacity of the storage pool during operation, and increase the capacity of the storage pool before it exceeds the threshold or becomes full.
-
Deleting some snapshots in the snapshot tree will not reduce storage pool usage because the space allocated as data in snapshot volumes in the storage pool will not be released. The unreleased space will be used again when storing new data in snapshot volumes in the same snapshot tree.
If you want to reduce usage of the storage pool, you need to delete all snapshots created in the corresponding snapshot tree.
-
Note that if the storage pool becomes full, and a snapshot failure occurs (snapshotStatus becomes "Error"), you cannot restore the snapshot even if you resolve the insufficient capacity by deleting volumes or adding drives.
If a failure occurs in the storage cluster (for example, as a result of stoppage of power to the storage cluster) while the write back mode with cache protection is disabled, data on the snapshot volume might be lost. When the write back mode with cache protection is enabled, data on the snapshot volume is protected. However, if a failure occurs in the storage cluster while failures exceeding the metadata redundancy for cache protection are occurring, the data on the snapshot volume is not protected even if the write back mode with cache protection is enabled.
Data in the snapshot volume (S-VOL or P/S-VOL) cannot be read or written if it is lost. Data in all snapshot volumes might be lost depending on the failure type. You have no choice but to delete snapshot volumes that cannot be read or written.
Back up your data to another device (e.g., via the host) as necessary.
Whether or not the operations can be performed for each snapshotStatus
The snapshot tree configuration and whether or not the operations can be performed for each snapshotStatus are as follows:
It might take about one minute for snapshotStatus to transit. To confirm the transition of snapshotStatus, wait about one minute and then check again.
■ Non-cascade configuration
"Snapshot operation for A" in the table indicates preparation for taking a snapshot of volume A or taking a snapshot of volume A. To prepare for taking a snapshot of volume B under volume A or take a snapshot of volume B under volume A, see "Snapshot operation for B" in "■ Cascade configuration."
snapshotType |
snapshotStatus of A |
Snapshot operation for A |
||||
---|---|---|---|---|---|---|
Prepare and take a snapshot (PrepareAndFinalize) |
Prepare |
Take a snapshot (Finalize) |
Restore |
Delete |
||
None |
None (A does not exist) |
〇 |
〇 |
- |
- |
- |
Snapshot |
Empty |
- |
- |
× |
× |
× * |
Preparing |
- |
- |
× |
× |
× |
|
Prepared |
- |
- |
〇 |
△ |
〇 |
|
Normal |
- |
- |
△ |
〇 |
〇 |
|
Restoring |
- |
- |
× |
× |
× |
|
Error |
- |
- |
× |
× |
〇 |
|
Deleting |
- |
- |
× |
× |
× |
|
* If removal processing is interrupted because the storage cluster has stopped, the processing can be completed by retrying the operation. 〇: Succeeds △: Succeeds but nothing is performed (The job ends normally without performing processing.) ×: Does not succeed -: The operation is not available |
■ Cascade configuration
You can take a snapshot of volume B only if the volume is operable.
snapshotType |
snapshotStatus of A |
Snapshot operation for B |
||||
---|---|---|---|---|---|---|
Prepare and take a snapshot (PrepareAndFinalize) |
Prepare |
Take a snapshot (Finalize) |
Restore |
Delete |
||
Snapshot |
Empty |
× |
× |
- |
- |
- |
Preparing |
× |
× |
- |
- |
- |
|
Prepared |
× |
〇 |
× |
△ |
〇 |
|
Normal |
〇 |
〇 |
〇 |
〇 |
〇 |
|
Restoring |
× |
× |
× |
△ |
〇 |
|
Error |
× |
× |
× |
× |
〇 |
|
Deleting |
× |
× |
- |
- |
- |
|
〇: Succeeds △: Succeeds but nothing is performed (The job ends normally without performing processing.) ×: Does not succeed -: The operation is not available |
You can take a snapshot of volume A only if the volume is operable.
snapshotType |
snapshotStatus of B |
Snapshot operation for A |
||||
---|---|---|---|---|---|---|
Prepare and take a snapshot (PrepareAndFinalize) |
Prepare |
Take a snapshot (Finalize) |
Restore |
Delete |
||
Snapshot |
Empty |
- |
- |
〇 |
〇 |
× |
Preparing |
- |
- |
〇 |
〇 |
× |
|
Prepared |
- |
- |
〇 |
〇 |
× |
|
Normal |
- |
- |
△ |
× |
× |
|
Restoring |
- |
- |
△ |
× |
× |
|
Error |
- |
- |
△ |
× |
× |
|
Deleting |
- |
- |
△ |
× |
× |
|
〇: Succeeds △: Succeeds but nothing is performed (The job ends normally without performing processing.) ×: Does not succeed -: The operation is not available |
■ When sharing the copy-source VOL among multiple snapshots
You can take a snapshot of volume B only if the volume is operable.
snapshotType |
snapshotStatus of A |
Snapshot operation for B |
||||
---|---|---|---|---|---|---|
Prepare and take a snapshot (PrepareAndFinalize) |
Prepare |
Take a snapshot (Finalize) |
Restore |
Delete |
||
Snapshot |
Empty |
〇 |
〇 |
〇 |
〇 |
〇 |
Preparing |
〇 |
〇 |
〇 |
〇 |
〇 |
|
Prepared |
〇 |
〇 |
〇 |
〇 |
〇 |
|
Normal |
〇 |
〇 |
〇 |
〇 |
〇 |
|
Restoring |
× |
〇 |
× |
× |
〇 |
|
Error |
× * |
× * |
〇 |
〇 |
〇 |
|
Deleting |
× |
× |
〇 |
〇 |
〇 |
|
〇: Succeeds △: Succeeds but nothing is performed (The job ends normally without performing processing.) ×: Does not succeed * Successful when the KARS06132-C message is output to the event log. |
■ QoS settings of snapshots
QoS settings can be applied to a P-VOL, S-VOL, and P/S-VOL.
For details about the QoS function, see About the Quality of Service (QoS) function.
Snapshots do not inherit QoS settings of the copy-source volumes.