ArcGIS Data Store 10.8.1 contains changes for the tile cache data store and spatiotemporal big data store.
Beginning with ArcGIS Data Store 10.8, you can define multiple backup locations for the tile cache data store and have multiple machines in a tile cache data store. You can also add a hosted feature layer or hosted spatiotemporal feature layer to a map in ArcGIS Pro and publish a map image layer. When you do, the map image layer references the data in the relational or spatiotemporal big data store, respectively.
See the 10.8 issues addressed list and 10.8.1 issues addressed list for information on fixes in the software.
New in ArcGIS Data Store 10.8.1
The following functionality is new in ArcGIS Data Store 10.8.1:
Configure a tile cache data store mode
A new operation on the configuredatastore utility at ArcGIS Data Store 10.8.1 allows you to choose the mode in which you deploy a tile cache data store—as primary and standby machines or a cluster containing multiple machines.
The multimachine cluster introduced at 10.8 allows you to scale the tile cache data store to accommodate large scene caches and large numbers of scene layers. However, accessing scene data in this mode can be slow when you use fewer than five machines in your tile cache data store. Therefore, at 10.8.1, you can use a primary and standby mode tile cache data store if you don't require scalability. In this mode, you can have a single machine (primary only) or two machines (primary and standby).
Tile cache data stores you create with 10.8.1 using the Data Store Configuration wizard or the configuredatastore utility default to the primary and standby mode. When you upgrade a tile cache data store, it maintains the mode it is in; therefore, if you upgrade a tile cache data store from a 10.7.1 or earlier release, your 10.8.1 tile cache data store mode will be primary and standby. If you upgrade from a 10.8 tile cache data store, the mode will be clustered. You can migrate from one mode to another in 10.8.1 using the configuredatastore utility. You can also change modes when restoring a tile cache data store to new machines.
Set a staging location that includes tile cache data store backups
ArcGIS Data Store extracts backup files from their compressed format and places them in a staging location before restoring relational or tile cache data stores. As tile cache data store use has increased, the size of the tile cache backup files has also increased. Therefore, when you use the changestaginglocation utility to change the staging location, the staging location is also used for tile cache data store backup files to avoid filling up the disk space on the tile cache data store machine.
Place the spatiotemporal big data store back in read-write mode
When your spatiotemporal big data store gets close to running out of disk space, ArcGIS Data Store places the data store in read-only mode to avoid data loss. When in read-only mode, no hosted spatiotemporal feature layers can be created and data streaming from ArcGIS GeoEvent Server cannot be archived. At 10.8.1, use the changedatastoremode utility to allow writes to the spatiotemporal big data store after you add disk space to the spatiotemporal big data store machines.
New in ArcGIS Data Store 10.8
The following sections describe functionality introduced in ArcGIS Data Store 10.8:
Scale the tile cache data store to meet scene layer demands
It is not uncommon to publish hosted scene layers that use a lot of disk space. In some cases, you also need to publish a large number of hosted scene layers. Two machines often do not suffice to store all these scene layer caches. To support scene layer scalability requirements, ArcGIS Data Store 10.8 allows you to create tile cache data stores that contain multiple machines. If you find your tile cache data store machines are nearing capacity and you need to publish more scene layers, you can add more machines to the tile cache data store. New scene layer caches will be stored on the new machines. Existing caches remain on the existing machines.
This new multimachine implementation means the following is now true for tile cache data stores:
- There are no longer primary and standby tile cache data store machines, and scene services are not restricted to reading caches from a single machine.
- You must open port 29079 in the firewalls of the tile cache data store machines to allow them to communicate with one another.
- For a highly available tile cache data store, deploy all the tile cache data store machines before you and your portal members publish hosted scene layers.
- You must rebalance the tile cache data store content in the following scenarios:
- You add machines to a tile cache data store that already contains scene caches and you want to redistribute the existing caches across all machines.
- A machine fails in the tile cache data store and you need to replace it.
See Add a machine to your data store for more information.
Define multiple backup locations for tile cache data stores
You can define multiple backup locations for the tile cache data store, including file shares, Amazon Simple Storage Service (S3) buckets, and Microsoft Azure Blob storage containers.
As a result of this change, the following is true for tile cache data stores:
- Newly created tile cache data stores do not have a default location; you must define a backup location before any backups can be created for a new tile cache data store.
- Before you upgrade a tile cache data store to ArcGIS Data Store 10.8, ensure the tile cache data store has a shared location set for the backup location. If it does not, you must register a new backup location for the tile cache data store after you upgrade, and the new location must be a shared location.
Map image layers can communicate directly with the relational or spatiotemporal big data store
Owners of hosted feature layers and hosted spatiotemporal feature layers in an ArcGIS Enterprise 10.8 portal can add one of these layers to a map in ArcGIS Pro 2.5 and publish a map image layer. This allows feature layer owners to make their hosted feature data available to others on a truly read-only basis. Once published, the map image layer connects directly to the data in the underlying data store: the relational data store if a hosted feature layer was in the map or the spatiotemporal big data store if a hosted spatiotemporal feature layer was in the map.
Improved publishing times for hosted feature layers
Various improvements were made to decrease the amount of time it takes to publish hosted feature layers when your relational data store already contains a lot of data.