Hi,
We're running a selfhosted DMS with self-hosted Cassandra and OpenSearch Cluster.
We were about to configure a backup of the DMA, just the DMA configuration and not the database, but it seems like that can't be done unless there's a backup for OpenSearch already set up.
https://docs.dataminer.services/user-guide/Advanced_Functionality/DataMiner_Agents/Backing_up_a_DataMiner_Agent.html
We wanted to use either options "without database". The purpose would be to restore a whole agent or restore individual config files if necessary.
It seems odd to me and I can't find any explanation as to why that is?
On a side note, there's not much written about what data OpenSearch actually stores.
Thank you for the answer. I never went further, since the Elastic search patch was red I assumed it was mandatory.
A couple of other things to mention:
According to your documentation
"If you select the Local paths option, all Agents in the Local paths list will automatically be selected, but you can clear the selection for some Agents if necessary. For the selected Agents, the backup package will be placed both in the default local folder C:\Skyline DataMiner\Backup and in the folder specified in the Local path box (on the same Agent)
You should clarify that it's only when pressing "Execute backup" that you get a dialogue box where you can deselect agents... I tried to deselect in the agents list but the boxes were all grayed out.
We would like to store the backup on another drive only and not use C:\Skyline DataMiner\Backup but it seems that is still not an option. There were once a workaround for that in your documentation but I can't seem to find it now. It involved links.
Hi Roger,
We are aware of the confusing UI in Cube related to the backup configuration.
The UI originates from the days that every DataMiner Agent had its local database(s). In a future version, the UI will be simplified as everything related to the databases today does no longer make sense in the context of back-up.
Today, we have self-hosted storage being CS & OpenSearch in clustered setup vs STaaS. With STaaS, backup should not be your concern and you can consider it being done automagically.
Self-hosted storage implies self-management and that includes backup.
So both storage options will exclude the databases from the 'DataMiner back-up' in the future and for now it's also advised to configure it as such.
In a self-hosted DataMiner system, trend data and element parameter data are stored in Cassandra, while alarms are stored in Elasticsearch.
Hope that helps and clarifies a bit.
Hi Roger,
When choosing for local backup paths, there is indeed no way to schedule the backups for a subset of the agents. And if you clear the custom paths for the local backups, they will fallback to C:Skyline DataMinerBackup. So the textboxes are there to specify another location for your local backups (for example another drive). When doing so, the backups should no longer be stored on the default location. In case they do, we'll have to look into the details of your DMS via a support ticket.
Manual triggers for backups via the button in Cube do indeed allow selection of the individual agents.