If you use Avamar (or NW) it will use its own mtree for all backups. If you use DDBoost agents (which are pretty much stripped down versions of what NW modules are in most cases), then you can create mtree for each category if you want and have it replicated any way you want. I personally still prefer to have single umbrella which has all backup metadata visible in one place therefore I use backup application (NW).
yes we use Avamar which will use its own mtree which is fine!.
The plan was to create an mtree based on data type.
We also need to consider data security. Example is multiple systems using the same mtree for backups (ie Oracle data). What stops a system from accessing other clients data? Is this via a different DDboost user or is it transparent security?
also need good links to some docs on designing this solution....
How would system A know what data of system B is there? Plus, system A is used by one user and system B by another, right? If this is oracle, so you use same user, you still access your own data only unless you can see data over RMAN catalog. If you wish to separate this further at level of DD, you can ofc use tenants or simply different ddboost user for different mtrees.
So from nw18.1 onwards using the Networker NMDA module and the Orchestrated Application Protection feature, one leverages the Dell BoostFS solution in combination with Networker to have Postgres write to a mounted DD filesystem.
If you only want to write to a DD using Postgres directly without a backup product, you can also leverage Dell BoostFS instead of mounting a DD share as CIFS or NFS to the Postgres system. Using BoostFS and the DDBoost protocol would result in faster backups as one would perform already client side (AKA source) deduplication instead of needing to send all backup data full-blown to a share and then have target side deduplication of data once it ends up on the DD system.
ble1
4 Operator
•
14.4K Posts
0
September 6th, 2017 14:00
If you use Avamar (or NW) it will use its own mtree for all backups. If you use DDBoost agents (which are pretty much stripped down versions of what NW modules are in most cases), then you can create mtree for each category if you want and have it replicated any way you want. I personally still prefer to have single umbrella which has all backup metadata visible in one place therefore I use backup application (NW).
mikelavery69
1 Rookie
•
47 Posts
0
September 7th, 2017 01:00
yes we use Avamar which will use its own mtree which is fine!.
The plan was to create an mtree based on data type.
We also need to consider data security. Example is multiple systems using the same mtree for backups (ie Oracle data). What stops a system from accessing other clients data? Is this via a different DDboost user or is it transparent security?
also need good links to some docs on designing this solution....
ble1
4 Operator
•
14.4K Posts
0
September 7th, 2017 11:00
How would system A know what data of system B is there? Plus, system A is used by one user and system B by another, right? If this is oracle, so you use same user, you still access your own data only unless you can see data over RMAN catalog. If you wish to separate this further at level of DD, you can ofc use tenants or simply different ddboost user for different mtrees.
mikelavery69
1 Rookie
•
47 Posts
0
September 11th, 2017 07:00
thanks
Based on DDOS 6.1 are there any BoostFS integration documents to cover MySQL, Oracle, MongoDB and MSSQL server?
vishwanath krishnappa
17 Posts
0
May 29th, 2023 23:00
Do Datadomain and DELL EMC have solution for Datadomain integration with Postgres SQL.
barry_beckers
393 Posts
0
June 19th, 2023 12:00
You mean standalone or rather integrated within a backup product like for example Dell Networker?
https://www.dell.com/support/kbdoc/en-uk/000158074?lang=en "NetWorker: How to configure a backup using NMDA 18.1 Orchestrated Application Protection feature?"
So from nw18.1 onwards using the Networker NMDA module and the Orchestrated Application Protection feature, one leverages the Dell BoostFS solution in combination with Networker to have Postgres write to a mounted DD filesystem.
The nw19.8 NMDA online manuals also states more about the Orchestrated Application Protection method: https://www.dell.com/support/manuals/en-uk/networker-module-for-databases-and-applications/nw_p_nmda_admin_guide_19.8/nmda-features-specific-to-orchestrated-application-protection?guid=guid-e3e59854-2826-4faf-8bb6-441f1cdfa051&lang=en-us supporting mongodb, mysql, postegresql and cassandra.
To install boostfs https://www.dell.com/support/manuals/en-uk/networker-module-for-databases-and-applications/nw_p_nmda_install/enabling-data-domain-boostfs-for-orchestrated-application-protection?guid=guid-a6d1045d-818e-4e3b-8bb4-40042c2d0533&lang=en-us for orchestrated application protection.
If you only want to write to a DD using Postgres directly without a backup product, you can also leverage Dell BoostFS instead of mounting a DD share as CIFS or NFS to the Postgres system. Using BoostFS and the DDBoost protocol would result in faster backups as one would perform already client side (AKA source) deduplication instead of needing to send all backup data full-blown to a share and then have target side deduplication of data once it ends up on the DD system.