General Main Release 10.3.0 CU11
Note
For known issues with this version, refer to Known issues.
Tip
For information on how to upgrade DataMiner, see Upgrading a DataMiner Agent.
Enhancements
Number of shards generated by Elasticsearch/OpenSearch has been optimized [ID 37619]
The number of shards generated by Elasticsearch and OpenSearch has now been optimized. These database will now generate less shards than before.
SNMPv3 responses now have a dynamic size [ID 37948]
Up to now, the size of SNMPv3 responses was limited to 16000 bytes. When larger responses were received, a timeout message would appear in StreamViewer.
From now on, the size of SNMPv3 responses will no longer be limited, meaning that the only constriction will be the maximum size of a UDP packet (i.e. 65000 bytes).
Note
When sending SNMPv3 messages, the size of those messages is still limited to 16000 bytes.
Reduction of number of information events when clients connect or disconnect [ID 37992]
In DataMiner Systems with a large number of Agents, up to now, each Agent in the DataMiner System would generate an information event when a client application connected to or disconnected from a particular Agent. From now on, when a client application connects to or disconnects from an Agent, only that particular Agent will generate an information event.
Enhanced performance when compiling QActions in SLScripting [ID 37993]
When QActions are compiled in SLScripting, several resources need to be loaded. Up to now, those resources would be loaded for every QAction individually. From now on, they will be loaded only once and stored in a cache in order to reduce memory and CPU overhead.
Every 10 seconds, resources that have not been referenced in the last 30 seconds will be removed from the cache.
Note
Uploading protocol or app packages that contain different versions of DLL files stored in the same location may not compile against the shipped version if the previous version was also uploaded and put in use in the previous 60 seconds. Forcing a recompilation of the QActions a minute later (causing the packages to be uploaded again) will yield the expected result.
Protocols and scripts should use NuGet packages as much as possible. The DLL files in those packages will then automatically be placed in folders by version. When protocols use custom non-NuGet DLL files, those should also be placed in folders by version.
With DLL files such as NewtonSoft, which protocols do not reference using NuGet, overwriting the DLL file with a newer version will cause protocols with QActions that have already been compiled to no longer work after a DataMiner restart as the correct strong-named assembly can no longer be found.
Elasticsearch/OpenSearch: TTL-based rollover for the CustomData and LoggerTables storage types is now disabled [ID 38000]
TTL-based rollover for the CustomData and LoggerTables storage types is now disabled.
SLAnalytics - Proactive cap detection: Cleanup of all proactive suggestion events at startup [ID 38004]
Up to now, after a restart of the SLAnalytics process, it could occur that a number of open proactive suggestion events were still left on the system. These would then remain open indefinitely unless a user would manually clear them.
From now on, all proactive suggestion events will be cleared when proactive cap detection is started, and trend alarm records will no longer be stored in the Cassandra database.
Security enhancements [ID 38040] [ID 38656]
A number of security enhancements have been made.
NATS: All nodes will now be considered primary nodes [ID 38089]
From now on, there will no longer be any primary and secondary NAS configurations. All nodes will now be considered primary and will be using their local credential files.
Also, when the NATS configuration is reset, the DMS IP addresses will now be collected via the online Failover agent.
Enhanced performance when deleting redundancy groups [ID 38173]
Because of a number of enhancements, overall performance has increased when deleting a redundancy group.
New BPA test 'Check Cluster SLNet Connections' [ID 38201] [ID 38208]
Run daily on every Agent in a DataMiner System, this new BPA test will trigger a local test on each Agent in the DMS that will
- check the connections between the different DMAs and between the DMAs in Failover setups, and
- report any communication problems.
For more information, see Check Cluster SLNet Connections.
SSH: Support for hmac-sha2-512-etm and hmac-sha2-256-etm [ID 38213]
DataMiner now supports two additional hash-based message authentication algorithms: hmac-sha2-512-etm and hmac-sha2-256-etm.
From now on, it will propose the following algorithms to the server in the following order:
- hmac-sha2-512-etm@openssh.com
- hmac-sha2-256-etm@openssh.com
- hmac-sha2-512
- hmac-sha2-256
- hmac-sha1
- hmac-md5
Fixes
Failover: Problems when using hostnames instead of virtual IP addresses [ID 32951] [ID 35380]
Up to now, a number of issues could occur when setting up a Failover system using hostnames instead of virtual IP addresses.
Problems with SLDataMiner [ID 37409]
SLDataMiner would leak memory when retrieving the baseline values of an element while the relative baseline value was being updated. Also, an error could occur in SLDataMiner after a service had been created, updated or deleted.
Apart from the above-mentioned fixes, memory management and overall error logging have also been improved.
Failover: Shared hostname would incorrectly always refer to the same agent when using gRPC [ID 37558]
On a Failover system with a shared hostname using gRPC connections, the shared hostname would incorrectly always refer to the same agent, whether it was online or offline. From now on, the shared hostname will always refer to the online agent.
SLDataGateway: Problem with casing when retrieving data from Elasticsearch/OpenSearch [ID 37835]
When SLDataGateway retrieved data from Elasticsearch/OpenSearch on behalf of a DataMiner app (e.g. Ticketing), in some cases, it would pass an incorrect result set to that app due to a casing issue.
Security: Users would incorrectly not be allowed to update the Visio file linked to a service [ID 37866]
Up to now, when users had the following permissions, they would not be allowed to update the Visio file associated with a service by using the Upload new visio file or Set new blank visio file commands:
- General > Visual Overview > Access Visual Overviews
- General > Visual Overview > Edit Visio drawings
- General > Services > Edit
- Write and Config permission for the service in question
From now on, when users have the permission to link a Visio file to a specific service, they will always be allowed to update the Visio file linked to that service.
Service & Resource Management: Timeout script in end event of booking would not get executed when the booking was set to end while the DMA was being stopped [ID 37911]
When the end event of a booking used a timeout script, in some cases, that script would not get executed when the booking was set to end while the DataMiner Agent was being stopped.
DataMiner Taskbar Utility: Problem when making SLTaskbarUtility perform an action using the command prompt [ID 37952]
When you made the DataMiner Taskbar Utility perform some action using the command prompt, the arguments would not be parsed correctly when no instance of SLTaskbarUtility was running.
MySQL: Problem when elements were started [ID 37973]
On systems with a MySQL database, in some cases, page handlers would be deleted too soon. This caused the following errors to be logged when an element was started:
Querying elementdata
Could not read the next page of element data for ...
Failed to query elementdata for parameter ...: General database failure.
Failed to query elementdata for ...: General database failure.
GQI: 'Could not find PK column' error after performing a query against an empty parameter table [ID 37978]
Up to now, in some rare cases, performing a GQI query against an empty parameter table would result in a Could not find PK column error. From now on, GQI will return an empty result set instead.
SLDataGateway would incorrectly keep waiting for acknowledgements from SLDataGatewayAPI [ID 37985]
SLDataGateway would incorrectly keep waiting for an acknowledgement from SLDataGatewayAPI, causing numerous Waiting for SLDataGatewayAPI to acknowledge ... entries to be added to the SLDataGateway.txt log file.
Problem with SLProtocol when a parameter was updated while an SLA window was changing [ID 37990]
When one of the following parameters was updated while an SLA window was changing, a run-time error could occur in SLProtocol:
- Base timestamp
- Monitor span
- Window size
- Window unit
- Window type
- Validity start
- Validity end
SLNet: Problem when a client application sent multiple messages to the same manager [ID 38025]
When a client application (e.g. DataMiner Cube) sent multiple messages to the same manager, in some cases, a number of those messages would return an exception and would not get processed.
Problem with protocol compliancies cache in SLNet [ID 38043]
The compliancies cache in SLNet, which keeps track of the minimum required version of a protocol or of whether a protocol supports Cassandra, was only refreshed when you uploaded a protocol.xml file that did not contain a <Compliancies> tag or when the cache was accessed for the first time.
GQI: Problem when retrieving a large amount of alarms [ID 38065]
When a GQI query had to retrieve a large amount of paged alarms, after a while, a timeout exception would be thrown even though none of the paged requests had timed out.
SLAnalytics: Problem when the trend icon calculation feature was disabled immediately after being enabled [ID 38072]
In some cases, SLAnalytics could partially get stuck when the trend icon calculation feature was disabled immediately after being enabled.
Failover: Problem with DVE elements and virtual function elements after a Failover switch [ID 38167]
After a Failover switch, in some cases, DVE elements or virtual function elements would not be loaded correctly. Also, new DVE elements would incorrectly not appear in the Surveyor when they were created while their parent element was hosted on the Failover setup that had switched.
Service & Resource Management: Incorrect trace data would be returned after performing a create, update or delete action using the ServiceManagerHelper [ID 38262]
When you retrieved the trace data after performing a create, update or delete action using the ServiceManagerHelper, in some cases, an error could be returned although the action that was performed had succeeded.
SLAnalytics could stop working when it lost its connection to SLNet during start-up [ID 38268]
Up to now, when SLAnalytics lost its connection to SLNet at a particular moment during start-up, it would stop working because it was not able to reach the database. From now on, when SLAnalytics loses its connection to SLNet at that particular moment during start-up, it will continue working and will try to connect to the database again as soon as its connection to SLNet has been re-established.
Correlation alarms with incorrect severity after a DataMiner restart [ID 38286]
After a DataMiner restart, in some cases, correlation alarms would have an incorrect severity.
BPA test 'Check Cluster SLNet Connections' could incorrectly report connection problems when it found a Failover system with a shared hostname [ID 38328]
Up to now, the BPA test Check Cluster SLNet Connections BPA could incorrectly report connection problems in a DataMiner System when it found a Failover setup with a shared hostname.
Failover: NATS would incorrectly be reconfigured when both agents were offline [ID 38349]
When both agents in a Failover setup were offline, in some cases, they would incorrectly reconfigure the NATS settings.
Problem with SLDataMiner when an enhanced service was not able to find some of its child services [ID 38583]
At DataMiner startup, SLDataMiner could throw an access violation exception when an enhanced service was not able to find some of its child services.