Reactivating the original vfiler

Boot the original filer and interrupt the boot process by pressing the Del or Esc key while the memory self-test is in progress.

Note
If you don't press the Del or Esc key in time, you can press Ctrl-C when prompted later during the boot; then choose option 5 (maintenance mode); then enter halt.

At the OK prompt, set the no-vfiler-ips? variable as follows:

setenv no-vfiler-ips? true

This ensures that the filer does not try to bind IP addresses already being used by the disaster-recovery vfiler. When the filer boots, the original vfiler will come up and report as running, but will not accept any read or write requests because all its interfaces will be unconfigured.

Resynchronize the mirrored volumes and qtrees.

For each volume and qtree owned by the original vfiler, enter the following command on the original filer:

snapmirror resync -S disaster_recovery_filer:/pathname 
production_filer:/pathname

Example

snapmirror resync -S drfiler:/vol/vfiler1/qtree1 prfiler:/vol/vfiler1/qtree

Troubleshooting

If the snapmirror resync command fails, telling you that there are no matching snapshots, you have probably accidentally deleted the snapshots that SnapMirror depends on. In this case you need to initialize SnapMirror using the snapmirror initialize command (see the na_snapmirror man page for details). Then continue with the next step of this procedure, but skip Step 7 when you get there.

Find out if the snapmirror.access option on the disaster-recovery filer is set to legacy. Enter the following command on the disaster-recovery filer:

options snapmirror 

If...

Then...

The options snapmirror command returns snapmirror.access legacy

Edit the file /etc/snapmirror.allow and add the host name of the original filer if it is not already there.

The options snapmirror command returns a list of host names, and the name of the original filer is not in the list

Use the options snapmirror command to add the host name of the original filer.

Example

options snapmirror.access host=fridge,toaster,prfiler

Run setup on the disaster-recovery vfiler and unconfigure its IP addresses.

Update the data on the original vfiler.

For each volume and qtree owned by the original vfiler, enter the following command on the original filer:

snapmirror update -S disaster_recovery_filer:/pathname 
production_filer:/pathname

Example

snapmirror update -S drfiler:/vol/vfiler1/qtree1 prfiler:/vol/vfiler1/qtree1

Stop SnapMirror transfers to the disaster-recovery vfiler.

For each volume and qtree owned by the original vfiler, enter the following command on the original filer:

snapmirror quiesce pathname 

Example

snapmirror quiesce /vol/vfiler1/qtree1

Note
This operation can take a long time. Use Ctrl-C to interrupt it if you need to.

Check that all the paths are quiesced:

snapmirror status

The status column in the output should show each path as Quiesced.

Break the SnapMirror relationship.

For each volume and qtree owned by the original vfiler, enter the following command on the original filer:

snapmirror break pathname 

Example

snapmirror break /vol/vfiler1/qtree1

On the original vfiler, run setup to configure the vfiler's IP addresses and configure the NIS and DNS servers.

Destroy the disaster-recovery vfiler. On the disaster-recovery filer, enter the following command:

vfiler destroy vfilername

vfilername is the name of the disaster-recovery vfiler.

Recreate the disaster-recovery vfiler.

From the disaster-recovery filer, perform the procedure Creating the disaster-recovery vfiler.

Boot the original filer and interrupt the boot process by pressing the Del or Esc key while the memory self-test is in progress.

Note
If you don't press the Del or Esc key in time, you can press Ctrl-C when prompted later during the boot; then choose option 5 (maintenance mode); then enter halt.

At the OK prompt, set the no-vfiler-ips? variable as follows:

setenv no-vfiler-ips? true

This ensures that the filer does not try to bind IP addresses already being used by the disaster-recovery vfiler.

At the OK prompt, enter boot.

Destroy the original vfiler. Enter the following command on the original filer:

vfiler destroy vfilername

vfilername is the name of the original vfiler.

Follow the procedure for Creating a replacement vfiler on the original filer or its replacement.

Boot the replacement filer.

Follow the procedure for Creating a replacement vfiler on the original filer or its replacement.

Step

Action

Prepare the new vfiler:

Perform the preparation steps under Preparing the destination filer.

Note
Remember that you are performing these checks to inventory the resources of the original filer or its replacement.

Create the new vfiler, following directions under Creating the vfiler.

Note
Perform those steps on the original filer or its replacement.

After performing those steps, you have a production vfiler and a backup vfiler, as you did before the disaster, but their roles are reversed: the vfiler you have just created is now the backup, and what was originally the disaster-recovery vfiler is still managing the live data. This may or may not suit your needs.

• If you want to put things back the way they were before the disaster, continue with the next step.
• If you don't care about the roles of the two vfilers, you are done.

Run setup on the disaster-recovery vfiler and unconfigure its IP addresses.

Update the data on the new vfiler.

For each volume and qtree owned by the new vfiler, enter the following command on the original filer or its replacement:

snapmirror update -S disaster_recovery_filer:/pathname 
production_filer:/pathname

Example

snapmirror update -S drfiler:/vol/vfiler1/qtree1 prfiler:/vol/vfiler1/qtree1

Activate the new vfiler, following directions under Activating the disaster-recovery vfiler.

Note
Perform these steps on the original filer or its replacement.

Destroy the disaster-recovery vfiler by entering the following command on the disaster-recovery hosting filer:

vfiler destroy vfilername

vfilername is the name of the original vfiler

Recreate the disaster-recovery vfiler.

From the original filer or its replacement, perform the procedure Creating the disaster-recovery vfiler.