Why does SnapMirror replication take a long time on my FSx for Netapp ONTAP volume?

Resolution

Slow SnapMirror transfer throughput creates long replication times. This might occur if you turned on the SnapMirror global throttle option. The SnapMirror global throttle option restricts the bandwidth used by incoming and outgoing SnapMirror and SnapVault transfers.

To turn off the global throttle feature, complete the following steps:

1.    To determine if the global throttle feature is turned on and to view the current throttle values, run the following command. In the command output, look for the replication.throttle.incoming.max_kbs and replication.throttle.outgoing.max_kbs values.

FsxIdxxxxxxx::> options

Example output

FsxIdxxxxxxx
replication.throttle.incoming.max_kbs 12500 -
replication.throttle.outgoing.max_kbs 12500 -

2.    To turn off the global throttle feature, run the following commands. The following example commands set the replication.throttle.incoming.max_kbs and replication.throttle.outgoing.max_kbs values to unlimited:

FsxIdxxxxxxx::> options -option-name replication.throttle.incoming.max_kbs unlimited
1 entry was modified.
FsxIdxxxxxxx::> options -option-name replication.throttle.outgoing.max_kbs unlimited
1 entry was modified.

3.    After you remove the throttle values, repeat step 1.

FsxIdxxxxxxx::> options

Example output

FsxIdxxxxxxx
replication.throttle.incoming.max_kbs unlimited - 
replication.throttle.outgoing.max_kbs unlimited -
2 entries were displayed.

FsxIdxxxxxxx::>

Note: You can keep the throttle values and manually control the transfer speed.

To set the maximum total bandwidth used by outgoing transfers on the source cluster, run the following command:

options -option-name replication.throttle.outgoing.max_kbs KBps

To set the maximum total bandwidth used by incoming transfers on the destination cluster, run the following command:

options -option-name replication.throttle.incoming.max_kbs KBps

Related information

Throttle SnapMirror globally on ONTAP 9 to reduce the effect on other workloads on the NetApp Knowledge Base website