New versions of Vortex Insight are released on a regular basis. This page lists all the fixed bugs and changes in each release of Vortex Insight.
You may also want to read Known Issues In Vortex Insight.
| Report ID.|| Description
|THM-1231|| Cloud and Fog 2.0 Support
Added support for the Cloud and Fog version 2.0.0.
The Cloud and Fog support is not backwards compatible and version 1.x is no longer supported in Insight.
|THM-1238|| QoS comparator does not account for presentation QoS.
The QoS compatibility comparison for readers and writers does not account for the presentation policy of the parent subscriber or publisher.
Solution: The presentation policy is included in the QoS comparison table.
|THM-1221|| Node Info and Process Info tabs do not display in Vortex Insight client browser using: Windows 10, Java 1.8.
If the NodeMonitor application on a given node is running on Windows 10 with Java 1.8, the Vortex Insight console starts
periodically reporting a SAXException regarding an unknown character reference found. This causes the Node Info and Process
Info tabs to not show up.
Fixed in Vortex OpenSplice 6.6.3p3 (OSPL-8563). The NodeMonitor implementation strips out the invalid form feed character
for the OS_INFO description information supplied by the Sigar library before publishing.
| Report ID.|| Description
|THM-1223|| The validation for the Node Statistics publication values in the Configuration Page does not work.
In the Vortex Insight's Configuration Page, under the "Node Monitor" tab, it was possible for the user to enter invalid values (any values other than
positive integer values) for the publishing rate of any node or process statistics property.
Solution: The issue was fixed by making sure Vortex Insight's client code checks the validity of the provided values and
strips out any invalid characters.
|THM-1220|| Debug logs still showing up in Vortex Insight console even when all log levels are set to INFO.
In the previous version of Vortex Insight, it was possible to see debug logs appear in Vortex Insight's console even
when all log levels were set to INFO.
Solution: The issue was fixed by updating the Vortex Insight to use a consistent logging mechanism
across the code base.
|THM-1212|| Update no_internals tree filter configuration to filter out new client side durability readers/writers.
New to Cafe applications, with every participant created, a d_historicalData reader and a d_historicalDataRequest writer
are created. These are considered built in entities and should be filtered out by the no_internals built in filter.
Solution: A new expression is added to the no_internals builtin filter configuration with the expression
|THM-1210|| Disabling any NodeInfo/Stat configuration setting from the Configuration Page does not work.
In the Vortex Insight's Configuration page, under the Node Monitor tab, attempting to disable reporting of any node
information or node statistics by publishing a value of zero for its input field did not work. Node information and
statistics would still be published.
Solution: The issue was fixed by updating Vortex Insight server code to properly interpret zero values as requests to
disable the property.
|THM-1208|| Unicast and Multicast locators do not show up in the Hops for Vortex Cloud/Fog in Vortex Insight.
In previous versions of Vortex Insight, unicast and multicast locator information was not reported in Vortex
Insight browser client even when published. In the Hop Profile view, the corresponding fields read "empty".
Solution: The issue was fixed by updating database calls so that information about published unicast and multicast
locators is retrieved along with the readers and writers. It will now show "empty" only when the information is not available.
| Report ID.|| Description
|THM-1033|| Insight doesn't recognize/group services in its browser
In previous versions of Insight, certain OpenSplice services such "networking bridge", "secure networking" and more would be displayed as user applications
in the domain tree browser instead of being shown as services (with participants under the "services" folder of the OpenSplice node). Additionally, Insight
would mis-represent these services as user applications if they were not created with default names. This is because Insight used to recognize services by
solely matching their names against a list of known default names.
Solution: Insight's service detection logic was improved to use internal information provided by OpenSplice that indicates if a given application is a service or not.
This approach guarantees accurate detection and representation of services in Insight.
|THM-1096|| Intermittent JdbcException occurring on Cloud domain disconnect
When Insight is connected to a Cloud domain, if the user requests Insight to disconnect from the domain while lots of cloud entity disposal
and restart updates are being received, it is possible a Jdbc exception occurs: "javax.persistence.PersistenceException: ERROR executing DML
bindLog error[Referential integrity constraint violation". After disconnect, DB table rows with garbage data remain.
Solution: In order to keep the Insight database in a consistent state, when a jdbcException is thrown on disconnect, the cloud DB tables
are cleaned up - ie all rows are deleted. An info statement is written to console, as well a debug statement with exception message
Disconnect continues as normal.
|THM-1097|| Time type QoS policies in Cloud monitoring display -1.-1 instead of "infinite".
When viewing QoS settings for cloud entity, QoS policies of type "Time" display the value "-1.-1" instead of "infinite".
This is because of some differing definitions of the infinite time constant int a Cloud context vs. an OSPL context.
So, the value "-1.-1" should be understood as meaning an "infinite" time duration value.
Solution: The value now correctly displays as "infinite".
|THM-1099|| Unable to discover new discovery or routing services when using tcp mode.
When Cloud services domain is configured using the tcp transport and Insight is connected to the cloud domain, new services on the domain
do not appear in the browser tree. Only disconnecting and re-connecting from the domain shows the new services.
Solution: Fixed in cloud tooling api.
From the "Threshold Monitor" tab under the Configuration page, and with the monitor configuration tables are empty,
if the user add a new participant monitor, then click Undo, the button does not react and the undo operation does not work.
This is because this particular scenario causes an error in the client browser.
|THM-1102|| Under certain network conditions, Insight server hangs if it tries to make a soap communication to an uninhabited port on a remote machine.
Under certain network conditions, Insight server would hang if it tried to communicate to a remote port that did not match the soap
service of the remote application it wished to communicate to. This happened for example, if Insight was originally connected to a remote application,
but then the application shutdown. Insight only knew that the entities of the application had been disposed and would try to communicate to the application to
do some cleanup. Normally, if there is no soap service on the receiving end, the communication request returns immediately, however, it was found that
under certain network conditions, the communication request would take minutes before returning. This was happening because there was no timeout mechanism in
Insight for soap communications.
Solution: The issue was fixed by introducing connection timeout properties in the soap communication layer of Insight. The default timeout for connection
is currently 5000ms meaning that Insight will abort a soap connection attempt if it takes more than that time; effectively preventing it to hang indefinitely in the call.
|THM-1103|| When domain updates happen, certain new publishers and subscribers are non-expandable in the domain tree view.
Whenever a new application starts up and depending on the order in which Insight receives updates about the new application,
it is possible for some publishers and subscribers of the new application to not be expandable in the domain tree view. This is
because Insight client tries to display some of the readers/writers of these publishers/subscribers while they are missing a label/name.
This is because Insight server communicates to client browser as soon as it receives data updates and without waiting to receive
updates about reader/writer's label.
Solution: Insight server has been updated to not send reader/writer data updates to the client browser until the reader/writer's
valid labels have been received.
|THM-1105|| DomainActor does not recover if crashed
The issue was previously incorrectly diagnosed: if a crash happens in the DomainActor, the actor will always recover as expected. However,
the domain tree browser of the client side will display duplicate tree entities. That is because the tree is not reset after the crash happens,
causing it to re-populate the same tree data twice.
Solution: This issue was resolved by having the domain tree browser notified about the crash. This causes the tree to clean up its content
and prepare for reloading it with fresh entity data from the server. A modal dialog is also displayed to notify the incident to the user.
|THM-1106|| The monitor graph visualization does not update properly after the associated ThresholdMonitorActor crashes
When viewing the threshold monitor graph about a monitored entity (topic or participant), the graph starts to update improperly if
for any reason, its associated ThresholdMonitorActor crashes. Although, the actor recovers properly, it seems that its internal state
is corrupted. In fact, this issue was caused because the actor scheduler was not being cleaned up when the actor crashed. This resulted
in two schedulers running concurrently against the ThresholdMonitorActor when it was restarted.
Solution: Insight source code has been updated so that the scheduler of the ThresholdMonitorActor is properly cleaned up whenever the
actor shuts down or crashes.
|THM-1107|| GlobalActor recovers if crashed, however user still has to disconnect and re-connect
The GlobalActor is the first internal process created by Insight. The GlobalActor globally manages the entire Insight server.
Currently, if this actor crashes, the client browser will no longer be able to reach the server. Although the GlobalActor process
recovers from any crashes, it does not autonomously re-connect to the domain that was being monitored. This means that the client
browser will lose connection to any monitored target (e.g. Vortex OpenSplice application) and will not be able to fulfill any
monitoring request. In the current implementation, if the above scenario occurs, the user is kicked back to the connection page,
with no feedback about what happened.
Solution: If the actor crash were to occur, the user will be redirected to the connect page, but now with a warning message indicating
that server restart has occurred.
|THM-1108|| Multiple clients do not get redirected to connect page if one client disconnects from Cloud domain.
If Insight is connected to Cloud, and there are multiple browser clients viewing, then if one client disconnects, all the
other clients remain on the cloud domain browser page, unable to see any entity profiles. The expected behaviour is for
all clients to be redirected to the connect page. This is how it is done when in normal domain connection mode.
Solution: Client behaviour for Vortex Cloud domain browsing fixed for multiple client connection. All clients get
redirected if one disconnects the server from the domain.
|THM-1120|| Wrong view of a simple federation.
Sometimes, depending on the way the built-in topic samples arrive and are processed by the Insight, it is possible to experience
an inconsistency in the domain tree view where the tree shows a federated system as a single process (e.g. showing an
application called "Opensplice Control Service"), before showing the proper federated system while not cleaning up the single
process application previously displayed.
Solution: This behavior was updated so that whenever Insight detects that a given system previously displayed as a single process is
actually a federated system, it automatically reloads the domain browser of all clients to reflect that new information and to
maintain the integrity of the tree's topology.
|THM-1154|| Vortex Insight now supports monitoring of Vortex OpenSplice 6.6.2, Vortex Cloud 1.3.1, Vortex Café 2.2.1, and Vortex Lite 2.0.2
|THM-1165|| Windows Start Menu shortcuts folder hierarchy.
The Windows start menu shortcuts for Vortex Insight have been reorganized. The top level shortcut folder is now
"VortexInsight V1.2.1". The actual installation paths in the file system are unchanged.
| Report ID.|| Description
|THM-488|| Threshold monitor time series graph timestamps are not accurate.
The timestamps in the time series graph are based on the current time in the client browser, which may
lead to slightly differing times on the graph plot from client to client.
Solution: Monitor time series graph now uses timestamps from the server side to display timeseries.
|THM-545|| FileNotFoundException occurring on server start if product install was performed as root.
If Vortex Insight is installed in Linux as the root user, the ./conf/generated.keystore file would be write
protected, causing an exception in the server if it is started as generic user.
Solution: The installer now sets write permissions to the conf directory.
|THM-624|| QOS Compatibility diagram doesn't automatically update from entity liveliness changes.
When viewing a DDS entity in the QoS Compatibility tab view, the diagram does redraw itself when the readers or
writers being viewed are no longer alive in the system, or if new relevant readers or writers are introduced.
Solution: The QoS Compatibility diagram now redraws itself when reader or writer liveliness changes.
|THM-633|| Displayed values for the threshold monitoring graph are not accurate when the participant or writer
When a given participant or all of a topic's writers leave the system, the threshold monitoring graph
does not accurately reflect what happens if we take into account the monitoring interval. What is observed
is that the graph plot goes straight to zero on the next sample received. The only time the value should go
straight to zero on the next graph point is if the interval is equal to the sampling rate. If the interval
is larger than the sampling rate, then when the samples start reporting zero writes, the graph should show
a decreasing slope, signifying the running average over the interval is decreasing.
Solution: The graph now display value changes with accurate rate of decrease.
|THM-642|| Vortex Insight displays not alive participants in the domain tree browser, even after they have departed
the from system.
It is possible sometimes for old applications and participants to remain behind in the tree browser after they
have been shut down. This is due to a case where the built in topic samples representing the liveliness of these
entities are not marked as not alive disposed, but as not alive no writers.
Solution: Insight now properly discards not alive participants, even when their samples not marked as disposed.
|THM-703|| Vortex Café applications in Linux show up inside a federation when they shouldn't.
Federations are a concept unique to Vortex OpenSplice deployments. Café applications should always be shown
as stand alone application entities.
Solution: The Café applications are no longer grouped in federations.
|THM-706|| Server should not start when licensing is invalid.
Currently, the license check is performed at domain connection time.
Solution: The license check has been moved to server start time. If no valid license is found, the Insight
server application will terminate and a invalid license message is printed to the console.
|THM-731|| Windows 8 - Uninstall menu item not available from tile.
In Windows 8, the right-click context menu for the Vortex Insight tile in the Metro dashboard UI is missing.
Solution: The uninstall menu item has been added to the Vortex Insight tile.
|THM-736|| Threshold monitor graph loading time depends on sampling rate.
If the sampling rate for a given monitor is quite long, then the user is presented with a loading graphic and
no data for the initial duration of the sampling rate.
Solution: Monitor graph now loads instantly and is no longer dependent on the sampling rate.
|THM-750|| Disconnecting from or shutting down Insight kills any insight_dds_connector processes running on the user's
In the case where multiple Insight servers are running to be connected to multiple domains via the
insight_dds_connector, then if one Insight server disconnects from its domain, all connector applications are
shut down, resulting in all other Insight servers to lose connectivity to their domains.
Solution: Disconnecting from or shutting down Insight server now kills only the insight_dds_connector child
process owned by that server.
|THM-760|| Writer/Reader rapidly published through Café gets placed under its Participant rather than its direct Pub/Sub
container in the Domain Tree.
In the case where a reader or writer's information becomes known before its parent publisher or subscriber becomes known,
the entities get placed under the participant as their direct parent temporarily until the publisher/subscriber info is
received. Sometimes though, due to a timing issue involving builtin topic sample arrival time and order, the reader and
writer parent references do not get updated.
Solution: Domain tree view in Insight now accurately maintains proper hierarchy between writers/readers
and publishers/subscribers no matter their arrival order.
|THM-858|| Vortex Lite applications showing under a Federation in Insight
Federations are a concept unique to Vortex OpenSplice deployments. Lite applications should always be shown
as stand alone application entities.
Solution: The Lite applications are no longer grouped in federations.
|THM-866|| Vortex Lite entities are not properly displayed in Insight.
Vortex Lite applications have no associated publisher or subscriber information, however a phantom publisher and
subscriber would appear in the domain tree browser under Lite participants. This results in confusion in where to
place reader and writer entities in the tree and often shows incorrect entity topology.
Solution: Insight now properly accounts for the lack of publisher and subscriber information for Vortex Lite entities.
Lite readers and writers are grouped under their parent participant, with no phantom publisher or subscriber.
|THM-936|| Reader and writer icons are missing from the topic tree view.
The pencil and glasses icons are missing in the topics tree view under the groupings by node.
Solution: The icons are restored in the tree.
|THM-941|| Threshold monitor page shows participant name instead of topic name in threshold config section when topic
threshold is "per topic"
Solution: The table value is changed to show the topic name, when the monitor is by topic.
|THM-973|| Latency statistics don't make sense - Min: 2147483648, Max: 0
In reader statistics, there are duration type fields for latency. The statistics values for these field do not make
much sense. In fact, the values displayed are meant to indicate that there are no updates to these stats, but meaning
of the data is not obvious in its raw form.
Solution: The latency statistics display has been redesigned to better reflect that min and max values have not been
|THM-986|| Insight does not read the application.conf file in the installdir/conf directory
In previous versions of Insight, the application.conf located in the installdir/conf directory was not read by Insight server. Additionally,
the user has to fiddle with a "logger.xml" file in order to change log levels in Insight.
Solution: the reference to the application.conf has been fixed and Insight server now reads configuration properties from the file.
(located in the installdir/conf directory). Additionally, a new "user_preferences.conf" (located in the installdir/conf directory as well)
has been added and the "logger.xml" file has been removed. The "user_preferences.conf" must be the user for any
configuration overrides (e.g. changing log levels, etc). The application.conf is not meant to be modified by the user, but it
references the "user_preferences.conf"
|THM-991|| Node monitor configuration is not published for * (all nodes)
In the previous version of Insight, if the user went to the "Node Monitor" tab under the Configuation page and published new node monitor
configurations to be applied to all nodes (by selecting the "All Nodes" option under the "Node" column in the configuration table), these
configurations will be ignored by Insight server even if the Insight client will give the impression that the changed went through.
Solution: The insight server has been fixed to properly recognize and process node monitoring configuration change requests for the specific case of
the "All Nodes" option.
|THM-992|| Fixed issue where the log message "Could not get statistics" is reported by Insight when user clicks on the Node to view statistics
In order to compute statistics for the Node, Insight server has to get and compute statistics for all descendants
of the Node.The issue happened when the Insight server was trying to get statistics about some of these descendant entities for which it had no cm object
reference. Cm object references are required in order to achieved retrieval of remote statistics information. However, in this case, the issue was that the
entities causing the error where "internal" DDS entities which do not have cm object nor do they have statistics.
Solution: The insight server has been fixed to not attempt to retrieve statistics for internal DDS entities or entities for which it has no cm object reference.
|THM-1003|| Implemented timeseries storage API for Insight metrics and statistics.
This is a new feature. Insight now allows user to record and store metrics and domain statistics as timeseries.
The feature is disabled by default. To enable this feature, switch the "opensplice.statistics.reporting.enabled" property to "true" in the "user_preferences.conf"
file located in the "installdir/conf" directory. Keep in mind you will need to install one of the Insight's supported reporting databases. InfluxDB is the one
|THM-1007|| Race condition during Insight disconnect-reconnect leaves left-over participants in Ospl (as reflected in insight's system browser tree)
Previously, if user disconnected and reconnected Insight very rapidly to the monitored OSPL target, an issue could occur where Insight
participants will not be cleaned up from the OSPL target. This was because in those scenarios, Insight will simply fail to free its participant
and some of its descendant entities due to the order in which the process was done.
Solution: The Insight server cleanup process was improved so that it always cleans up its participants and descendants even if situations where
rapid disconenct-reconnect operations occur.
|THM-1008|| Vortex Platform Windows installer does not set script variables if drive letter is not specified
This issue occured if for example, user tried to install Insigth on windows using the unattended mode feature.
THe installer will install the product, but fail to properly set Insight script variables. As a result, user will not be able to start Insight.
Solution: Insight windows installer workflow was improved to also work for cases where the install path does not contain a drive letter, and make sure
that the values of all required script variables are always properly set.
|THM-1017|| NPE when selecting Network reader
In the Insight's previous version, if the user connected insight to an ospl target, click on the target's networking reader in the domain tree and then
on the "QOS Settings" tab, the settings would not display properly and a null pointer error would occur in Insight's console. This is because the Insight
server was trying to get certain QOS settings which are not available for the networking reader (cause they were not applicable to that entity).
Solution: The issue has been fixed by improving the way Insight server handles these received QOS by making it more robust. The null pointer issue does
not happen anymore.
|THM-1073|| Input text fields on insight system page do not respond to "left" and "right" keystrokes when user is viewing statistics
In previous versions of Insight, if the user clicked on an entity on the domain tree to view statistics, then went to any input text field on the page and
tried to enter some text then use the "left" and "right" keyboard strokes to navigate within the input field, that navigation will not work. The insight client
will not respond to those keystrokes when done from within any input field on the page. This was because the cubism engine used to display change rate graph
for statistics will override handling of these keystrokes and prevent default behavior.
Solution: To fix this issue, a check has been added to the cubism keydown.context handling function so that it does not prevent default event handling of
keystrokes when the focus of the page is on an input text field.
|THM-1089|| Hyperlink error in the singlehtml user guide.
In previous versions of Insight, in the html user guide (the one accessible from within the tool itself), the sidebar link to section "14.5. Configuration" will
take the user instead to section "13.1. Configuration."
This was because both sections had the same div ids (derived from their identical names).
Solution: The issue was fixed by giving different title names for the conflicting sections in their respective .rst files
|THM-1098|| If Insight is connected in Cloud mode, and other browser clients join in, they only see the domain browser view forever stuck loading
If the server was in Cloud connection mode, the standard route (eg. http://localhost:9000/) will not be redirected to the Cloud view (http://localhost:9000/cloud).
As result, the Insight page would load in the Entreprise page and show the "Loading..." message indefinitely in the tree view
Solution: The issue has been fixed by having Insight server detect which mode it is on (Cloud vs. Entreprise) and execute a re-direct based on that.
If the current connection session is cloud, then the "/" route now redirects to "/cloud". If the current connection session is normal, then the "/cloud"
route redirects to "/" route.