Table of Contents

General Feature Release 10.3.3

Note

For known issues with this version, refer to Known issues.

Tip

Highlights

OpenSearch & Amazon OpenSearch Service [ID_34651]

It is now possible to install a dedicated OpenSearch indexing database cluster as an alternative for Elasticsearch. This indexing cluster also requires a Cassandra cluster.

At present, all OpenSearch versions are supported:

  • 1.X
  • 2.X

For more information on setup and configuration, see OpenSearch database.

Note

It is also possible to use Amazon OpenSearch Service on AWS as an alternative to an on-premises hosted Elasticsearch/OpenSearch cluster. For more information on setup and configuration, see Amazon OpenSearch Service.

Amazon Keyspaces [ID_34872]

It is now possible to use the Amazon Keyspaces Service on AWS as an alternative for a Cassandra Cluster setup.

Note
  • Amazon Keyspaces does not support all Cassandra functionality, most notably indices on columns. As a result, some queries against logger tables (including SLAs) may be slower on Amazon Keyspaces compared to Cassandra.
  • The only consistency level supported is Local Quorum. See Supported Apache Cassandra consistency levels in Amazon Keyspaces. No matter the configuration, the consistency level will always be overwritten to Local Quorum.
  • The only replication strategy supported is the Amazon Keyspaces specific Single-Region strategy, which is not configurable.
  • Migrating existing databases to Amazon Keyspaces is not yet supported.

For more information on setup and configuration, see Amazon Keyspaces Service.

Other features

DataMiner Object Model: New 'ClientReadOnly' and 'AllowMultipleSections' properties [ID_35172]

Next to a ReadOnly property, the DomStatusFieldDescriptorLink class now also has a ClientReadOnly property.

  • ReadOnly determines whether values of the field in question are read-only when the field has a particular status.

    When a field is marked as read-only for a specified status, the values cannot be changed when the DomInstance has this status. This also means that if no values were present before transitioning to this status, no values can be added as long as the DomInstance continues to have this status.

  • ClientReadOnly determines whether users are allowed to assign a value to the field in question in the UI.

    Unlike the ReadOnly property, this property does allow users to assign a value to the field using the API (e.g. in a script).

If the ReadOnly property is true, the value of ClientReadOnly is ignored.

An AllowMultipleSections property has now been added to the SectionDefinitionLink (non-status system) and DomStatusSectionDefinitionLink (status system) classes. This property will allow you to define whether a DomInstance can have multiple Sections for a particular SectionDefinition.

When multiple sections are added to a DomInstance, and this is not marked as allowed, a DomInstanceError will be returned in the TraceData with the MultipleSectionsNotAllowedForSectionDefinition error reason. Additionally, in the InvalidSections property of the error, you will be able to find the ID(s) of the SectionDefinition(s) of which too many sections were found.

Rules that apply with regard to multiple sections:

  • Non-status system (DomDefinition and SectionDefinitionLinks):

    • Multiple Sections for the same SectionDefinition are allowed on a DomInstance if the SectionDefinition has a link on the DomDefinition that has the AllowMultipleSections property set to true.

    • When a Section is added to an existing DomInstance that contains a ReadOnly field (marked as such in the FieldDescriptor), then that field cannot be given a value in the UI. It can only be given a value via the API or a script.

    • There are no limitations with regard to removing Sections.

  • Status system (DomBehaviorDefinition & DomStatusSectionDefinitionLinks):

    • Multiple Sections for the same SectionDefinition are allowed in a specific status on one DomInstance if, for that status, the SectionDefinition has a link on the DomBehaviorDefinition that has the AllowMultipleSections property set to true.

    • When a Section is added, any field marked as ReadOnly or ClientReadOnly will not be assignable from the UI. However, if they are only marked as ClientReadOnly, they will be assignable via the API.

    • Removing a Section is not possible when a field of the section in question is marked as ReadOnly on the link. When this behavior is required, but you still want to prevent users from assigning a value in the UI, use the new ClientReadOnly property instead.

Important

This is a breaking change. Previously, it was possible to add multiple sections without having to set any property. After upgrading to DataMiner 10.3.3/10.3.0, you will need to update any existing DomDefinitions with multiple Sections.

Each section now has to contain all required fields

Up to now, if there were multiple sections, it was allowed for some of those sections to not contain all the required fields. From now on, every section must contain each and every required field.

DataMiner Object Models: Action buttons can now be configured to launch an interactive Automation script when clicked [ID_35226]

An action button can now be configured to launch an interactive Automation script when clicked. To do so, set the IsInteractive property of the action to true.

When such a button is clicked in a low-code app, the UI of the interactive Automation script will be displayed in a pop-up window on top of the low-code app.

Note

One button can only execute one action. So, one button can only execute one interactive Automation script.

DataMiner Object Models: New field descriptors [ID_35278]

Two new field descriptors have been added to the DataMiner Object Models:

  • GroupFieldDescriptor: Can be used to define that a field should contain the name of a DataMiner user group.

  • UserFieldDescriptor: Can be used to define that a field should contain the name of a DataMiner user. There is a GroupNames property that can be used to define which groups the user can be a part of.

SLLogCollector will now order the Standalone BPA Executor tool to execute all BPA tests available in the system [ID_35436]

Each time the SLLogCollector tool is run, it will now order the Standalone BPA Executor tool to execute all BPA tests available in the system and store the results in the C:\Skyline DataMiner\Logging\WatchDog\Reports\Pending Reports folder.

The names of the files containing the test results will have the following format: <BPA Name>_<Date(yyyy-MM-dd_HH)>

Changes

Enhancements

Security enhancements [ID_34894] [ID_35331]

A number of security enhancements have been made.

Elasticsearch: 'Request Entity Too Large (413)' errors will now be prevented when writing data in bulk [ID_34937]

When data was written to Elasticsearch in bulk, up to now, DataMiner would throw a Request Entity Too Large (413) error when the amount of data exceeded the 100 MB limit.

From now on, DataMiner will detect when too much data is being sent in a single request and will split up the data into smaller parts.

SLLogCollector: Multiple instances can now be run simultaneously [ID_35204]

Multiple instances of the SLLogCollector tool can now be run simultaneously.

Cassandra Cluster: Enhanced query performance [ID_35247]

Previously, queries against Cassandra Cluster would always be executed with a page size of 1000, even when the limit defined in the query was smaller than 1000. This resulted in excess data being returned. From now on, the page size will be adjusted according to the limit defined in the query if it is lower than the default page size.

This change will considerably improve overall query performance, especially when retrieving trend data.

Note

This change will not enhance performance when requesting trend data for an element that has no trend data points before the requested window. In cases like this, the full two-year range of shards will be queried to try and find an initial point.

SLAnalytics - Pattern matching: Manually created tags will now be saved as pattern occurrences [ID_35299]

From now on, when you define a tag for pattern matching, the pattern you selected will be saved as a pattern occurrence in the Elasticsearch database and highlighted in bright orange, similar to so-called "streaming matches", which are detected while tracking for trend patterns whenever a trended parameter is updated.

In the SLDataGateway process, a number of memory enhancements have been made with regard to the management of average trend data.

SLAnalytics - Behavioral anomaly detection: Suggestion events and alarm events for a DVE child element will now be generated on that same DVE child element [ID_35332]

When a behavioral anomaly was detected on a DVE child element, up to now, the suggestion event or the alarm event would be generated on the parent element. From now on, it will be generated on the child element instead.

Maps: Markers will now move more gradual when zooming [ID_35337]

Because of a number of enhancements, markers will now move more gradual when zooming.

Alarm templates - Smart baseline calculations: NullReferenceException prevented & enhanced exception logging [ID_35348]

In some cases, a Baseline Calculation Failed: System.NullReferenceException: Object reference not set to an instance of an object error would be added to the SLSmartBaselineManager.txt log file. The issue causing that error has now been fixed.

Also, log entries indicating an exception thrown during baseline calculations will now include details regarding the element and parameter associated with the exception.

Maps: Enhanced zooming behavior [ID_35351]

From now on, when you zoom in or out, the data of the previous zoom level will stay visible until the data of the current zoom level has been loaded.

DataMiner upgrade: Installation of Microsoft .NET 6.0 [ID_35363]

During a DataMiner upgrade, Microsoft .NET 6.0 will now be installed if not installed already.

SLSNMPAgent log entries will now include the alarm ID [ID_35404]

When an entry is added to the SLSNMPAgent.txt log file, in most cases, that entry will now include the alarm ID.

Example:

  • Old format: Received ACK from SNMP Manager SNMP - LFR

  • New format: Received ACK from SNMP Manager SNMP - LFR for alarm 239/4270232

SLAnalytics - Automatic incident tracking: Enhanced performance when fetching relation information [ID_35414]

Because of a number of enhancements, overall performance has increased when fetching relation information for the automatic incident tracking feature.

Service & Resource Management: Enhanced performance when changing active function files [ID_35424]

Because of a number of enhancements, overall performance has increased when changing an active function file.

Also, in the Cube UI, users will receive more concise feedback regarding the impact of the change. Up to now, they would receive a list of all items affected by the change. From now on, the list of affected items will only show up to 10 affected items per object type.

SLAnalytics - Behavioral anomaly detection: No longer available for discrete parameters [ID_35465]

From now on, anomaly detection will no longer be available for discrete parameters.

Automation: Enhanced memory usage [ID_35502]

Because of a number of enhancements, overall memory usage of the SLAutomation process has improved.

Fixes

Cassandra Cluster: Every DMA would incorrectly try to delete any possible old Cassandra compaction and repair tasks found in the entire DMS [ID_31923]

At start-up, every DataMiner Agent with a Cassandra Cluster configuration would incorrectly try to delete any possible old Cassandra compaction and repair tasks found in the entire DMS.

From now on, at start-up, every DataMiner Agent with a Cassandra Cluster configuration will only delete the old Cassandra compaction and repair tasks found locally.

DataMiner Taskbar Utility: Problem when stopping DataMiner [ID_34790]

Up to now, when you right-clicked the DataMiner Taskbar Utility system tray icon and selected Stop DataMiner while keeping the SHIFT button pressed, the SLWatchdog process would incorrectly also be stopped. In a Failover setup, this would prevent the backup agent from acquiring the virtual IP address.

Also, after DataMiner had been stopped, up to now, the SLXml, SLLog and SLDataGateway processes would incorrectly start up again.

Problem with SLLog when logging large entries regarding failed Elasticsearch query requests/responses [ID_35037]

Up to now, an error could occur in SLLog when adding large entries regarding failed Elasticsearch query requests/responses.

When a direct view table was updated, the wrong columns could be updated in the source element [ID_35075]

When a direct view table was updated while one of the source elements was stopped, due to a column translation issue, the wrong columns could be updated in that source element the moment it was started again.

SLDataGateway would leak memory when offloading average trend data for DVE elements [ID_35167]

The SLDataGateway process would leak memory when offloading average trend data for DVE elements.

Service & Resource Management: Setting a new function file to active would incorrectly not cause the function DVEs of elements using a production version of the protocol to be updated [ID_35178]

When a new function file was set to active, up to now, the function DVEs of elements using a production version of the protocol in question would incorrectly not be updated.

Problem with CassandraBackup.exe when the configuration file of a general database of type 'Cassandra' had a DBServer element that contained multiple host addresses [ID_35253]

When, in the configuration file of a general database of type "Cassandra", the <DBServer> element contained multiple host addresses, up to now, the CassandraBackup.exe tool would parse those host addresses incorrectly and would not take any database backup.

From now on, when the <DBServer> element contains multiple host addresses including that of the local database, the CassandraBackup.exe tool will take a backup of the local database.

SLDataGateway could end up with an excessive number of HealthMonitor.Refresh threads [ID_35286]

In some cases, the SLDataGateway process could end up with an excessive number of HealthMonitor.Refresh threads.

Some agents in the cluster would incorrectly remove the run-time hosting agent info they had stored for another agent [ID_35287]

When run-time connections were being set up between agents or when a midnight sync was being executed, some agents in the DataMiner cluster would incorrectly remove the run-time hosting agent information they had stored for another agent in the same cluster.

DataMiner Object Models: Permission checks for DOM modules requiring view permission 'None' were too strict [ID_35305]

If a DOM module is created without specifying SecuritySettings, the view permission is set to "None".

Up to now, the check to determine whether a user had the view permission set to "None", would only return true for the Administrator or users in the Administrator group. From now on, when the required view permission is "None", permission checks will no longer be performed.

External authentication via SAML: Issues fixed when using Okta as identity provider [ID_35374]

Using Okta as identity provider, it would incorrectly no longer be possible to read out signed assertions. Also, when the group claim setting is enabled in the DataMiner.xml file, the user will now be added to the correct groups.

Up to now, in case of a claim mismatch, an exception would be thrown. From now on, an entry containing a clear message will be added to the SLNet.txt log file instead.

Problem when connecting to Azure AD via a proxy [ID_35382]

When a DataMiner Agent connected to Azure Active Directory was sealed off from the internet but had access to a proxy, users would not be able to log in due to SLDataMiner failing to use the proxy to connect to Azure Active Directory.

In case a DataMiner Agent has to connect to Azure AD via a proxy, then that proxy has to be configured by means of the following netsh command:

netsh winhttp set proxy <proxyaddress> <bypasslist>

Cassandra: TTL setting of spectrum trace data would not be applied correctly [ID_35385]

In a Cassandra database, the "time to live" (TTL) setting of spectrum trace data would not be applied correctly. As a result, this type of data would never be removed.

Problem with SLElement when stopping an EPM element [ID_35439]

When an EPM element was stopped, in some rare cases, an error could occur in SLElement.

SLDataGateway would not correctly return errors when querying SLA logger tables in a Cassandra Cluster [ID_35440]

SLDataGateway would not correctly return errors when querying SLA logger tables in a Cassandra Cluster, causing an error to occur in SLProtocol.

Problem when sending northbound SNMP inform messages in chronological order [ID_35441]

When northbound SNMP inform messages were being sent in chronological order, an error could occur when sending those messages suddenly stopped.

Automation: 'engine.RunClientProgram' overload with two parameters would incorrectly always be run synchronously [ID_35476]

An engine.RunClientProgram overload with two parameters, of which the second one controls whether the method is run either synchronously or asynchronously, would incorrectly always be run synchronously.

RunClientProgram(String applicationPath, bool waitForCompletion)

Business Intelligence: Problem when correcting outages on an SLA with a week-based window [ID_35503]

When outages on an SLA with a week-based window were corrected, the History Statistics Table that started in the first 9 weeks of every year would incorrectly not get updated.

SLAnalytics - Behavioral anomaly detection: Two identical behavioral anomaly alarms would incorrectly be created [ID_35511]

In some cases, two identical behavioral anomaly alarms would incorrectly be created.

SLAnalytics : Problem after a DVE parent element had been deleted [ID_35521]

In some cases, an error could occur in the SLAnalytics process after a DVE parent element had been deleted.

Cassandra Cluster: Rows would incorrectly be added without TTL value [ID_35789]

When a row was added to a Cassandra Cluster table, no TTL value would be inserted into the TTL column of that row.

Propose changes