Disaster Recovery - Fail over from an accessible primary volume
If a failure occurs on the primary site, and performance is degrading while the primary storage is still accessible, customers can reroute their operations to the secondary site by initiating a failover.
Before you start the failover, make sure that the required host-authorization is in place.
Authorized hosts and volumes must be in the same data center. For example, you can't have a replica volume in London and the host in Amsterdam. Both must be in London or both must be in Amsterdam.
Authorizing the host
Before you begin, make sure that the host that is to access the Block Storage for Classic volume is authorized. For more information, see Authorizing the host in the consoleAuthorizing the host from the CLIAuthorizing the host with Terraform.
Starting a failover from a volume to its replica
If a failure event is imminent, you can start an Immediate failover or a "Controlled Failover" to your destination, or target, volume.
When you choose an Immediate Failover, the last successfully replicated snapshot is activated, and the volume is made available for mounting. The target volume becomes active in less time compared to a Controlled Failover. However, any data that was written to the source volume since the previous replication cycle is lost.
A Controlled Failover is the best choice when you want to test the failover function. It's also the best option when it’s more important to continue operations at the replica location with the most recent data. In a Controlled Failover, a new snapshot is taken and copied over to the replica location. After the data is successfully copied over, the volume is made available for mounting.
When a failover is started, the replication relationship is flipped. Your target volume becomes your source volume, and your former source volume becomes your target as indicated by the Volume Name followed by REP.
Before you proceed with these steps, disconnect the volume. Failure to do so, results in corruption and data loss.
Fail over to replica in the console
Failovers are started under Storage > Block Storage for Classic in the IBM Cloud® console.
-
Click your active LUN (“source”).
-
Click Replica, and click Actions.
-
Select Controlled Failover or Immediate Failover.
Expect a message across the page that states that the failover is in progress. Additionally, an icon appears next to your volume on the Block Storage for Classic page that indicates that an active transaction is occurring. Hovering over the icon produces a window that shows the transaction. The icon disappears when the transaction is complete. During the failover process, configuration-related actions are read-only. You can't edit any snapshot schedule or change snapshot space. The event is logged in replication history. When your target volume is live, you get another message. Your original source volume's LUN Name updates to end in "REP" and its Status becomes Inactive.
-
Click View All Block Storage for Classic.
-
Click your active LUN (it was your previous target volume).
-
Mount and attach your storage volume to the host. For more information, see Connecting your storage.
Fail over to replica from the CLI
Before you begin, decide on the CLI client that you want to use.
- You can either install the IBM Cloud CLI and install the SL plug-in with
ibmcloud plugin install sl
. For more information, see Extending IBM Cloud CLI with plug-ins. - Or, you can install the SLCLI.
Initiating a failover from the IBMCLOUD CLI
You can use the ibmcloud sl block replica-failover
command to fail over operations from the source volume to the replica volume. The following example initiates a failover from the source share 560156918
to the replica
share 560382016
.
$ ibmcloud sl block replica-failover 560156918 560382016
OK
Failover of volume 560156918 to replica 560382016 is now in progress.
For more information about all of the parameters that are available for this command, see ibmcloud sl block replica-failover.
Initiating a failover from the SLCLI
To fail over a block volume to a specific replicant volume, use the following command.
$ slcli block replica-failover --help
Usage: slcli block replica-failover [OPTIONS] VOLUME_ID
Options:
--replicant-id TEXT ID of the replicant volume
--immediate Failover to replicant immediately.
-h, --help Show this message and exit.
During the failover process, configuration-related actions are read-only. You can't edit any snapshot schedule or change snapshot space. The event is logged in replication history. When your target volume is live, you get another message. Your original source volume's Status becomes Inactive.
Mount and attach your storage volume to the host. For more information, see Connecting your storage.
Starting a failback from a volume to its replica
When your original source volume is repaired, you can start a controlled Failback to your original source volume. In a controlled Failback,
- The acting source volume is taken offline.
- A snapshot is taken.
- The replication cycle is completed.
- The just-taken data snapshot is activated.
- The source volume becomes active for mounting.
When a Failback is started, the replication relationship is flipped again. Your source volume is restored as your source volume, and your target volume is the target volume again as indicated by the LUN Name followed by REP.
Failback in the console
Failbacks are started under Storage, Block Storage for Classic in the IBM Cloud® console.
-
Click your active LUN ("target").
-
Next, click Replica, and click Actions.
-
Select Failback.
Expect a message across the page that shows that the failover is in progress. Additionally, an icon appears next to your volume on the Block Storage for Classic that indicates that an active transaction is occurring. Hovering over the icon produces a window that shows the transaction. The icon disappears when the transaction is complete. During the Failback process, configuration-related actions are read-only. You can't edit any snapshot schedule or change snapshot space. The event is logged in replication history.
-
Next, click View All Block Storage for Classic.
-
Click your active LUN ("source").
-
Mount and attach your storage volume to the host. For more information, see Connecting your storage.
Failing back from the CLI
Initiating a failback from the IBMCLOUDCLI
You can use the ibmcloud sl block replica-failback
command to fail back operations from the replica volume to the original source volume. The following example initiates a failback to the original source share 560156918
.
$ ibmcloud sl block replica-failback 560156918
OK
Failback of volume 560156918 is now in progress.
For more information about all of the parameters that are available for this command, see ibmcloud sl block replica-failback.
Initiating a failback from the SL CLI
To fail back a block volume from a specific replicant volume.
$ slcli block replica-failback --help
Usage: slcli block replica-failback [OPTIONS] VOLUME_ID
Options:
--replicant-id TEXT ID of the replicant volume
-h, --help Show this message and exit.
Mount and attach your storage volume to the host. For more information, see Connecting your storage.