Standard Mode Environments Deployments Configuration¶

Dialog to configure deployments on a standard environment¶

The common options that you can configure for standard mode environments are:

• Enable deployments: Only environments with this option enabled admit working as the target of a deployment. The deployment will fail otherwise.

• Save backup in VCS when the deployment finishes: If selected, a backup of the deployment will be saved in the Version Control System you have configured. The backup will be performed both for successful and unsuccessful deployments, using the metadata of the first enabled server in the first enabled cluster.

• Rollback when the deployment fails: When the deployment fails in a server, the changes are rolled back. Note that if the deployment was successful in previous servers, those servers are not rolled back.

• Deployment type: There are two types of deployments according to whether the Solution Manager interacts with the load balancer or not.

  • In the Simple deployment mode, the Solution Manager only updates the metadata of the servers, one by one, without any interaction with the load balancer. This option is intended for offline applications.

  • In the Without service interruption deployment mode, the Solution Manager disables Virtual DataPort servers or clusters in the load balancer and enables them again when the deployment finishes. To deal with the load balancer, the Solution Manager uses the scripts defined in the Deployment Scripts section.

When the deployment is configured as Without service interruption, you can choose among different strategies to perform the promotion. The actual options available depend on the topology of your infrastructure, which is shown by the parameter Number of clusters installed. This parameter shows the number of enabled clusters that have at least one enabled Virtual DataPort server in the environment.

In those environments with only one enabled cluster, the subsection Promotion policy when the environment contains only one cluster will be available.

• VDP Servers use a shared metadata database: Specifies if the Virtual DataPort servers have an external database configured to store the metadata.

  • No: each Virtual DataPort server virtual machine has its own metadata database.

  • Yes: the Denodo Virtual DataPort servers share an external database for the metadata.

  Warning

  Notice that the Without service interruption deployment type is only available if the environment has more than one cluster if shared metadata is enabled.

Configuration options available when the environment contains one cluster¶

In such scenarios, you can configure the following options:

• Server promotion strategy: Solution Manager uses the scripts Enable server in the load balancer and Disable server in the load balancer to enable and disable Virtual DataPort servers in the load balancer respectively (disabled Virtual DataPort servers are ignored in this process). It can promote servers in two different ways:

  • The Half strategy disables half of the Virtual DataPort servers in the load balancer first, performs the deployment and enables them again. Then, it repeats the process with the other half. This option ensures that the potential system inconsistencies will only last a short time.

  • The One By One strategy disables the first Virtual DataPort server of the cluster (according to the order in the catalog tree of the Solution Manager) in the load balancer, performs the deployment and enables it again. This process is repeated server by server, so the system may present inconsistencies for a longer time than for the Half strategy.

• Secondary shared cache available: To avoid any problem with the cache of Virtual DataPort servers, Solution Manager allows you to define two different cache data sources between which to alternate on every deployment. As servers update their metadata during the deployment, their caches swap from the current (old) data source to the other one, which contains updated data and, maybe, views with updated schemas. In this way, those servers that have not been updated yet are not affected by the changes in the cache. For this to work, there have to be two different schemas in the same database that can operate as cache for Virtual DataPort.

• First cache URL: URL of one of the schemas that can work as the cache for the Virtual DataPort servers.

• Second cache URL: URL of the other schema that can work as the cache for the Virtual DataPort servers.

In those environments composed by several enabled clusters, the subsection Promotion policy when the environment contains more than one cluster will be available.

Configuration options available when the environment contains more than one cluster¶

This subsection only contains the option Cluster promotion strategy, which configures how clusters are disabled and enabled in the load balancer during the promotion. Take into account that the Solution Manager uses the scripts Enable cluster in the load balancer and Disable cluster in the load balancer, respectively, for this purpose (disabled clusters are ignored in this process). There are two strategies available:

• The Half strategy disables half of the clusters in the load balancer first, performs the deployment for the servers defined in those clusters and enables them again. Then, it repeats the process with the other half. This option ensures that the potential system inconsistencies will only last a short time.

• The One By One strategy disables the first cluster of the environment (according to the order in the catalog tree of the Solution Manager) in the load balancer, performs the deployment for the servers it contains and enables it again. This process is repeated cluster by cluster, so the system may present inconsistencies for a long time.

Cloud Environments Deployments Configuration¶

Dialog to configure deployments on a cloud environment¶

The common options that you can configure for cloud environments are:

• Enable deployments: Only environments with this option enabled admit working as the target of a deployment. The deployment will fail otherwise.

• Minimize Downtime: Specifies if the servers provide service or not during cluster recreation needed due to deployment changes.

  • No: the Solution Manager recreates the cluster without waiting for load balancer health checks to complete for the new servers. The old servers are removed after adding the new ones, so there is a time when the servers are not providing service. This option is intended for offline applications.

  • Yes: the Solution Manager recreates the cluster waiting for load balancer health checks to complete for the new servers and then removes the old servers. This option minimizes possible servers downtime.

• Deployment type: Specifies if the Solution Manager will perform changes directly in the current servers or in a new temporary server.

  • Update the image and recreate the cluster: With this option enabled, Solution Manager will launch a new virtual machine, perform changes there, create an image using that virtual machine as base and recreate the cluster applying minimize downtime or not depending to configuration. The cluster capacity will not be affected during the deployment and no changes will be performed in it if something goes wrong, but deployments can take several minutes due to image creation and cluster recreation.

  Important

  This option is not available if the Virtual DataPort servers in the target cluster use a shared metadata database since metadata are not stored in the instances and there is no need to create an image.

  • Perform changes directly in the servers: In this case, Solution Manager performs the changes directly in the current servers of the cluster.

  Warning

  This option is usually the fastest one but the capacity of the cluster may be affected during the deployment and changes may need to be rolled back manually if something goes wrong. Notice that if the Virtual Data Port cluster has auto-scaling enabled, then this option is only available when the Virtual DataPort cluster uses a shared metadata database.

  Important

  Keep in mind that since this option does not create and update the image in the cluster, if you do a recreate cluster operation after doing a deployment of this type you might lose the changes.

• VDP Servers use a shared metadata database: Specifies if the Virtual DataPort servers have an external database configured to store the metadata.

  • No: each Virtual DataPort server virtual machine has its own metadata database.

  • Yes: the Denodo Virtual DataPort servers share an external database for the metadata.

  Warning

  This option only supports the Perform changes directly in the servers deployment type, in this case performing the changes only in the first enabled server of the cluster.

  Important

  When shared metadata database option is enabled, this option has priority over deployment type, so changes will always be performed in one server and no image will be created nor the cluster will be recreated.

Data Catalog Server Synchronization¶

It is also possible to configure Solution Manager to synchronize Data Catalog servers for all environment types.

Data Catalog Server Synchronization configuration¶

The options that you can configure for Data Catalog Synchronization are:

• Enable Data Catalog synchronization: Data Catalog servers will be synchronized only if this option is enabled.

• Select the Data Catalog elements to synchronize: Select the type of elements to synchronize:

  • Views: Synchronize views metadata.

  • Web services: Synchronize web services metadata.

  • All: Synchronize metadata of all previous types.

• Data Catalog synchronize priority: After synchronizing, the schemas will be updated to reflect the current ones of the server, and the descriptions can be updated in three different ways:

  • Server: the descriptions will be overwritten with the server ones.

  • Local: all the descriptions of the Data Catalog will be preserved.

  • Server with local changes: the descriptions will be overwritten with the server ones, except the ones that have been edited from the Data Catalog (that will be preserved).

• Select the Data Catalog clusters to synchronize: Select the cluster(s) with Data Catalog servers to synchronize. Only clusters with at least one enabled Data Catalog server will be synchronized.

• Logical server name to synchronize: Optional. You can specify a logical name of a server to synchronize (this corresponds to the name of a Virtual DataPort server in Data Catalog). You only need to configure this if in the Data Catalog server there is more than one Virtual DataPort server.

  Warning

  If logical server name to synchronize field is empty, the metadata for all the Virtual DataPort servers defined in the Data Catalog server will be synchronized.