When implementing replication over slow links or recovering replication after a long outage the initial synchronization or re-synchronization can be very time consuming and impact overall performance of the deduplication solution.
By seeding we put in place the data on the secondary server using alternative transport mechanisms to pre-load it before we start replication since the data is already there the synchronization will be a lot faster as no data needs to be transferred.
Seeding is only possible for simple A->B replication
The ADE components must be in the same (identical pathnames) location on source and destination
Moving the data needs to be done OFF-LINE hence the seeding will require downtime.
The procedure requires editing critical configuration items that can make the Deduplication solution unusable and potentially lead to data-loss, the engagement of an Atempo Consultant is highly recommended.
On the Source server:
Run the ade_admin command bck_list and record the output
Typically this will be similar to what is shown below
STREAM=/HSS/Streams * if HSS is enabled
ALN=/AtempoDedupEngine/default/ALN_HVDS * if ALN or HVDS is enabled
If the destination is already installed and replication has been used
Using ade_admin perform an del_all_repli is required to remove the current replication relationship
Using ade_admin disable the garbage collector and compactor
Hence ade_admin service disable 7 and service disable 9
Now you should shutdown the source server
On the destination server:
If ADE is already installed :
Use ade_admin info command to collect the Server UUID of the installation
Output from the info command will be similar to :
Server version : 188.8.131.5252 (started 3 days, 21 hours and 20 minutes ago)
Server UUID : 82367175-2d39-8c42-b25c-7befa3175465
Hostname : ESC-HSS:23232
Verify that the ade_admin bck_list command shows the same output than what has been seen on the source server if this is not the case you will need to perform an un-install / install of the destination server software.
Copy the data :
With both ADE services down copy/transfer the contents of the paths listed by the bck_list command to the destination server overwriting the current content.
This will result in the destination server to be a CLONE of the source so make sure the source is offline when bringing up the destination.
Change the uuid on the destination
If you have a previous uuid use the ade_admin change_uuid command
If you have a new installation use the ade_admin renew_uuid command
This will stop the service and change its identity
Make sure both source and destination are up and running.
Use ade_admin or the GUI to re-establish the replication
Expl: ade_admin add_repli destination:port
Post seeding :
After both servers got in sync (allow for 24h) you should re-enable the garbage collector and compactor using ade_admin service enable 7 and service enable 9.
Articles / Posts
- Advisory ID: MIRIA-2021-0001
- Log4J / Log4Shell vulnerabilities
- Advisory ID : LINA/ADE-2021-0001
- How Miria distributes Jobs to Datamovers
- Best practices to increase the level of security ADE/LINA
- Restore Miria PostgreSQL DB – Windows
- Restore Miria PostgreSQL DB – Linux
- How to start the catalog maintenance task
- Tina Schedule the catalog maintenance task
- Atempo Licence Manager setup fails on debian 10
- Lina – EN (15)
- Miria – EN (6)
- Tina – EN (13)