SnapMirror creates snapshots on the source volume, which are copied to the mirror. The snapshot name is in the following format:
filer_volume.number
filer is the host name of the destination filer.
volume is the name of the destination volume (the mirror).
number is the number for the snapshot starting at 1.
SnapMirror automatically deletes old snapshots that are no longer necessary for data replication.
The following example describes the snapshots that are created on the source volume and copied to the mirror.
In this example, data is replicated from /vol/vol1of filerA to /vol/vol2 of filerB.
To create a baseline version of a mirror, filerA creates a snapshot file named /vol/vol1/filerB_vol2.1 on filerA. All snapshots in /vol/vol1 of filerA, including /vol/vol1/filerB_vol2.1, are transferred to /vol/vol2 of filerB.
For example, the snap list command on filerB generates the following display after /vol/vol1/filerB_vol2.1 is transferred from filerA to filerB.
Volume vol2
working.....
%/used %/total date name
-------- -------- ------------ --------
0% ( 0%) 0% ( 0%) Nov 17 10:50 filerB_vol2.1
1% ( 0%) 0% ( 0%) Nov 17 10:00 hourly.0
1% ( 0%) 0% ( 0%) Nov 17 00:00 nightly.0
1% ( 0%) 0% ( 0%) Nov 15 16:00 hourly.1
1% ( 0%) 1% ( 0%) Nov 15 15:00 hourly.2
2% ( 0%) 1% ( 0%) Nov 15 14:00 hourly.3
2% ( 0%) 1% ( 0%) Nov 15 13:00 hourly.4
2% ( 0%) 1% ( 0%) Nov 15 12:00 hourly.5
When it is time to update the mirror, another snapshot is created on filerA. The snap list command on filerA generates the following display after /vol/vol1/filerB_vol2.2 is created on filerA:
Volume vol1
working....
%/used %/total date name
-------- ---------- ------------ --------
0% ( 0%) 0% ( 0%) Nov 17 10:52 filerB_vol2.2 (busy)
0% ( 0%) 0% ( 0%) Nov 17 10:51 filerB_vol2.1
1% ( 0%) 0% ( 0%) Nov 17 10:00 hourly.0
1% ( 0%) 0% ( 0%) Nov 17 00:00 nightly.0
1% ( 0%) 0% ( 0%) Nov 15 16:00 hourly.1
1% ( 0%) 1% ( 0%) Nov 15 15:00 hourly.2
After filerB_vol2.2 is transferred from filerA to filerB, both the filerB_vol2.1 and filerB_vol2.2 snapshots exist on filerB.
On filerA, however, filerB_vol2.1 is no longer needed and is deleted; only filerB_vol2.2 remains.
The filer fails to replicate data if it cannot find the required snapshots in the source and destination volumes. This section provides an example illustrating what happens if you inadvertently delete a snapshot required for data replication.
In this example, data is replicated from /vol/vol1 of filerA to /vol/vol2 of filerB. Suppose you use the snap delete command on filerA to delete the filerB_vol2.2 snapshot. When filerB tried to update the mirror according to the schedule in the snapmirror.conf file, it could not find the snapshot to determine what incremental changes are required in the mirror.
As a result, filerB displays the following error message.
The source filer requires a complete transfer. The destination volume is online and must be offline for a complete transfer.
FilerA displays the following error message:
Destination filerB could not accept a complete transfer.
These messages mean that filerB tries to create a baseline version of the mirror, but doing so requires that the mirror on filerB be taken off-line.
To proceed, use the vol offline command to take the mirror off-line and reboot the filer. At the next scheduled mirror update, the filer creates a baseline version of the mirror.
Copyright © 1999, Network Appliance, Inc. All rights reserved.
Part No. 210-02334
File: mirror3.htm
Last updated: 02/24/99 12:17:14