Performing CTE maintenance with an LDT GuardPoint in a Pacemaker Setup

Unguard the LDT GuardPoint with Pacemaker setup

1. Stop the SQL server, type:

   sudo systemctl stop mssql-server

   Copy

2. Delete the existing resource constraints from LDT and the colocation from the Pacemaker setup.

   Example

   A. If it exists, get the ID of the PCS constraint co-location. In the example below the, ID is: co-location_set_msmsmt

   sudo pcs constraint show --full
   .
   .
   .
   Resource Sets:
   set mssql_fs mssql_gpfs mssql_gpldt sequential=false (id:co-location_set_msmsmt_set) setoptions score=INFINITY (id:co-location_set_msmsmt)

   Copy

   B. Remove the PCS constraint co-location and PCS resource, type:

   sudo pcs constraint remove co-location_set_msmsmt
   sudo pcs resource delete mssql-ldtmgp

   Copy

   Expected Result: Current LDT GuardPoint status should be now be: unguarded.

   Example

   secfsd -status guard

   Copy

   Response

   GuardPoint          Policy                          Type    ConfigState  Status       Reason
   ----------          ------                          ----    -----------  ------       ------
   /var/opt/mssql/LDT  LDT_clear_to_CS1_aes256_onhost  manual  unguarded    not guarded  Inactive

   Copy

3. Unguard the LDT GuardPoint from the Cluster Host Group in CipherTrust Manager.

   Expected Result: The LDT GuardPoint should no longer be displayed when secfsd -status guard is executed.

Perform CTE maintenance

Now you can perform CipherTrust Transparent Encryption maintenance tasks that require stopping secfs such as:

• Stop CipherTrust Transparent Encryption

• Upgrade CipherTrust Transparent Encryption

• Uninstall CipherTrust Transparent Encryption

• Create a new baseline database for LDT on the same LDT &{gp}

See CTE Agent for Linux Advanced Configuration for more information.

Note

Before creating a new baseline database for LDT: Ensure that your target LDT GuardPoint path does not have an ldt xattr value on it so that the directory can trigger the initial rekey at the start of guarding. If it does have an ldt xattr value from a previous setup, then use the command voradmin ldt xattr delete <LDT GuardPoint> to remove the old xattr value.

Re-Enable LDT GuardPoint with Pacemaker Setup

1. Re-enable the LDT GuardPoint from the Cluster Host Group in CipherTrust Manager. Wait until the GuardPoint is visible again and in the unguarded state before moving to the next step below.

   Example

   # secfsd -status guard

   Copy

   Response

   GuardPoint          Policy                          Type    ConfigState  Status       Reason
   ----------          ------                          ----    -----------  ------       ------
   /var/opt/mssql/LDT  LDT_clear_to_CS1_aes256_onhost  manual  unguarded    not guarded  Inactive

   Copy

2. Add the resource mssql-ldtmgp back into the Pacemaker setting and verify that it has started:

   Example

   sudo pcs resource create mssql-ldtmgp ocf:heartbeat:mgp mgpdir=/var/opt/mssql/LDT --group Apache_Grp-fs

   Copy

   The LDT GuardPoint should now be automatically active again in Pacemaker.

   # secfsd -status guard

   Copy

   Response

   GuardPoint          Policy                          Type    ConfigState  Status   Reason
   ----------          ------                          ----    -----------  ------   ------
   /var/opt/mssql/LDT  LDT_clear_to_CS1_aes256_onhost  manual  guarded      guarded  N/A

   Copy

3. Start SQL server, type:

   sudo systemctl start mssql-server

   Copy

4. Add the LDT resource constraints and co-location back into the Pacemaker setting.

   Example

   A. Add failover resource constraints back:

   sudo pcs constraint order start mssql_gpldt then start mssql-ldtmgp
   
   sudo pcs constraint order stop mssql-ldtmgp then stop mssql_gpldt

   Copy

   B. If constraint co-location was used, add it back:

   sudo pcs constraint co-location set mssql_fs mssql_gpfs mssql_gpldt sequential=false

   Copy

5. Restart Pacemaker so that it can pick up the newly created resource configurations.

   A. Stop SQL server, type:

   sudo systemctl stop mssql-server

   Copy

   B. Reboot Pacemaker, type:

   sudo systemctl restart pacemaker

   Copy

   C. Re-start SQL server, type:

   sudo systemctl start mssql-server

   Copy