Table of Contents

General Main Release 10.5.0 CU18 - Preview

Important

We are still working on this release. Some release notes may still be modified or moved to a later release. Check back soon for updates!

Important

Before you upgrade to this DataMiner version:

Tip

Changes

Enhancements

Security enhancements [ID 45646]

A number of security enhancements have been made.

APIGateway: gRPC connections that go through the Azure Cloud Relay service will now buffer event messages [ID 45671]

From now on, gRPC connections that go through the Azure Cloud Relay service will buffer event messages until the client confirms they have been received.

This will allow those connections to survive a temporary outage of the Azure Cloud Relay service, for example when restarting or deploying a new version.

ConfigureIIS.bat script will now ensure a dedicated Application Pool for the API application [ID 45842]

The ConfigureIIS.bat script will now ensure a dedicated Application Pool for the API application. This will avoid needless restarts of the API when files running under the same DefaultAppPool would change inside WebPages.

This new application pool is called DataMiner WebAPI AppPool. it is solely intended to serve as pool for the web API, and will not recycle periodically.

Fixes

Problem with SLPort when sending a large message over a WebSocket connection [ID 45625]

Up to now, when a large message (e.g., a message with a size of 400 KB) was sent over a WebSocket connection, in some cases, an internal buffer issue could cause the SLPort process to stop unexpectedly.

Protocol object outside of QAction run would incorrectly not be notified when the element was stopped [ID 45749]

When an SLProtocol or SLProtocolExt object that was passed as argument in the QAction method had been stored and reused when the QAction was already finished, up to now, this object would not be notified when the element was stopped.

This could lead to issues. For example, when a separate thread that was running in SLScripting when the element was stopped made calls on the SLProtocol object, the SLScripting process could crash as the connection with the SLProtocol process was no longer valid.

Also, calling protocol.IsActive would incorrectly indicate that the element was still active.

Problem when using the 'Get parameter table by alias' data source against a STaaS database [ID 45766]

The Get parameter table by alias data source retrieves a parameter table from the indexing database using the specified alias.

Up to now, this data source would only check whether the DataMiner System included an indexing database. It would not check the type of the database. As it currently only supports Elasticsearch and OpenSearch, up to now, exceptions would be thrown when it was used to retrieve data from a STaaS database.

From now on, the Get parameter table by alias data source will only be available when the DataMiner System includes an indexing database of type Elasticsearch or OpenSearch.

Problem occurring while SL* services were being shut down would prevent DataMiner from starting up again [ID 45839]

Up to now, while the SL* services were being shut down, in some cases, an access violation crash could occur.

As a result, DataMiner could fail to start up again.

Reconnecting a WMI connection could cause the SLProtocol process to stop unexpectedly [ID 45851]

Up to now, in some cases, reconnecting a WMI connection could cause the SLProtocol process to stop unexpectedly.

In addition, opening StreamViewer would incorrectly show all items in the tree structure as Undefined.

From now on, reconnecting a WMI connection will no longer cause SLProtocol to stop unexpectedly, and StreamViewer will correctly show the group and action executing the WMI query.

Problem with SLProtocol when a queued QAction finished after an element had been stopped [ID 45882]

Up to now, when an element was stopped while queued QActions were still running, in some cases, one of those QActions could finish after SLProtocol had already cleaned up its internal objects.

As a result, when that QAction thread then tried to update element metrics, it would attempt to access an object that had already been deleted, causing the SLProtocol process to crash.

Smart-serial client connection state incorrectly shown as undefined [ID 45931]

Up to now, when an element with a smart-serial connection acted as a client, in some cases, the Connection State column in the Communication Info table on the General parameters page would incorrectly show Undefined.

From now on, that column will correctly show the actual connection state, e.g., Connected.

Problem with SLDataMiner when SLWatchdog requested element statistics while an element was being stopped [ID 45945]

In some rare cases, SLDataMiner could stop unexpectedly when SLWatchdog requested statistics about the number of active elements while an element was being stopped.

SLSNMPManager process could stop working unexpectedly when it received a malformed SNMP packet [ID 45993]

Up to now, the SLSNMPManager process could stop working unexpectedly when, while using SNMP++, it received a malformed SNMP packet containing an integer type with length zero.

BrokerGateway could stop unexpectedly on startup because of concurrent semaphore access [ID 46009]

In some cases, BrokerGateway could stop unexpectedly during startup when multiple threads tried to create a semaphore with the same key at the same time. This could cause the semaphore to be released more than once, which made BrokerGateway crash.

Invalid matrix 'columntypes' definition could cause SLProtocol to stop unexpectedly [ID 46011]

Up to now, when a matrix parameter had fewer columntypes defined in its options than there were dimensions, SLProtocol could stop unexpectedly when protocol.SendToDisplay was called on that matrix parameter.

From now on, missing matrix outputs that are not covered by columntypes will be handled correctly.

Service card reports could show incorrect alarm counts on clustered databases [ID 46048]

Up to now, on clustered databases (Cassandra Cluster or STaaS), alarm counts in the Reports section of a service card could be duplicated by the number of DataMiner Agents involved in that service.

From now on, those alarm counts will be calculated correctly when a service includes elements hosted on multiple Agents.

Improve this page
Advanced search syntax