RTView® Enterprise Monitor |
|
Known Limitations
24854: Broker HWM values are sometimes wrong or showing NaN in Capacity Table
Setting the History Rows to a very low value (i.e. 10) can cause erratic fluctuations in HWM values, due to insufficient data from which to calculate the HWM.
Deployment
21725: Historian log reports autocommit error in MYSQL and MSSQL and Oracle
Users of the Historian may experience the following error when using MySQL: MYSQL: SQLException: Can't call commit when autocommit=true Or the following error when using Microsoft SQL Server: MSSQL: SQLException: commit() should not be called while in auto-commit mode. This error is benign and does not indicate any loss of data. For the sake of clean logs, it can be addressed in the following manner: MySQL - Relax autocommit in the URL specified for your connection, as follows: jdbc:mysql://MyHOST:3306/MyDB?relaxAutoCommit=true Microsoft SQL Server - Use the current drivers provided by Microsoft: sqljdbc42.jar
General
20157: History Table Statistics does not calculate Deltas until one update cycle passed
The History Table Statistics display will initially show incorrect data for Delta Rows. You must wait one update cycle for this to be resolved.
24859: HTML displays don't always populate after restart
Occasionally the HTML displays do not show new cache data, especially while a data server has recently come online. At this time the solution is to manually refresh the browser page.
Scripts
20838: "stop_rtv all" does not include a server listed after "viewer/builder" lines
Documentation clarification: The rtvservers.dat file contains groups of servers called a "config", associated by a common name, e.g. "central". Such groups must be contiguous in the file. That is to say, a config may not be split up in the file; all the servers it includes must appear contiguously.
Solution Package for Apache Kafka
24865: Each Kafka consumer must have a unique client id
RTView requires each Kafka consumer to have a unique client id in order to properly index the consumer data in the displays. This value is assigned at startup via Kafka properties or command-line argument. The Kafka documentation says: client.id An id string to pass to the server when making requests. The purpose of this is to be able to track the source of requests beyond just ip/port by allowing a logical application name to be included in server-side request logging.
Solution Package for Docker
21559: Unable to remove a stopped container in Docker: 'device or resource busy'
If cAdvisor is started after other containers, it prevents Docker from removing those containers. If cAdvisor is the first container loaded into Docker, then other containers can be removed without any issue. The workaround is to start the cAdvisor before any other container.
Solution Package for IBM Websphere Application Server
20205: "Response Time" KM for WSM mapped to non-history column
The Response Time key metric for WebSphere (CI Type WAS-APP) is not supported in the Key Metrics History displays at this time.
Solution Package for MySQL
21923: MysqlServerProperties cache columns don't support mixed MySQL versions
The MysqlServerProperties cache columns are created based on the first connection property that returns result. If subsequent connection results return that are a different version of MySQL, they may contain a different order or number of columns. The result is that the MysqlServerProperties cache may show row values in the wrong columns. This problem only appears if you look at the cache directly for troubleshooting purposes -- it does not effect any displays.
Solution Package for Oracle Coherence
18662: Multi-cluster configuration does not support mixed Coherence versions
Different versions of Coherence cannot be reliably monitored via a multi-cluster configuration, due to different MBean structures of Coherence versions. The first JMX connection that is established will set the MBean structure used by RTView, which can block any subsequent connections that use a different structure.
Solution Package for Solace PubSub+
21865: Calculated Totals on the Neighbor Summary display are not a direct sum
The calculated totals field in the Solace Monitor "Neighbor Summary" display are not always an exact sum of the above calculated fields. This is because the values in the Totals row are not derived by adding up the Data, Control, and Discards values. These calculations are derived from a direct query to the Solace appliance of the total number of messages to its CSPF Neighbors. The delta and rate calculations are therefor not always exact sum of the calculations against the Data, Control, and Discarded messages.
23672: RTViewSolaceMonitor - "only new projects can be configured" warning incorrectly displays when old 4.2 war files are in use
When an old RTViewSolaceMonitor 4.2 project is ran against 5.x and the user has forgotten to update war files in Tomcat, loading the config UI will result in the "only new projects can be configured" warning.
24750: Multiple broker alerts do not support a delay of 0 seconds
Setting the alert delay to 0 seconds can result in a false positive trigger for some multiple broker alerts (e.g SolGuaranteedMsgingNoMsgSpoolAdActive).
24775: Polling rate of Stats Pump connections not currently configurable
The configurable Poll Rates fields in the Configuration Application under Solution Package Configuration > Solace > Data Collection do not apply to data collected via Enterprise Stats (referred to also as the statspump data source).
24793: solace event module regex errors for client flow events from brokers previous to 9.4.0.78
Solace changed the format of the CLIENT_CLIENT_OPEN_FLOW and CLIENT_CLIENT_CLOSE_FLOW messages in 9.4.0.78. These events from brokers prior to 9.4.0.78 will generate a regex parsing error and the additional fields will not be parsed correctly. These messages will still be available in the SolEventModuleEvents and SolEventModuleAlert caches but the Additional Fields value for those events will be blank.
24794: v8.11 VMR brokers do not show Platform information
Solace VMR brokers v8.11 do not current display Platform information in the Provisioning data.
24798: SolGuaranteedMsgingMatePortDown triggers for standalone brokers
The SolGuaranteedMsgingMatePortDown alert triggers for standalone brokers where Guaranteed Messaging has been enable. We recommend that users disable this alert for standalone brokers.
24801: SolGuaranteedMsgingMatePortDown does not clear until failover occurs
The SolGuaranteedMsgingMatePortDown alert will not clear until a failover occurs.
24804: # Endpoints refer to both active and expired instances
A number of the items in displays do not correctly reflect when an item has been expired due to dynamic removal of the connection without restarting the dataserver. Some examples of this are: - The Top 10 endpoints includes expired endpoints without any change in color - The Connected icons in the Brokers table do not turn red when a connection is no longer available, and the "Show Connected Only" checkbox does not filter them out
24827: Regex parsing errors in soleventmodule.log
The following events may produce errors in the solaceeventmodule.log if the logging level is set to warn or lower: CLIENT_CLIENT_CONNECT - regex parsing does not support commas in capabilities field CLIENT_CLIENT_CONNECT_FAIL - regex id does not handle case where client-username field is missing VPN_SOLCACHE_LOST_MSG_STATE - regex parsing does not support commas in description field SYSTEM_CLUSTERING_LINK_DOWN - regex does not support missing 'reason' field
24829: Link Up status does not support Solace Software Brokers
The Link Up column on the Interface Summary page currently does not currently provide values for Solace Software Brokers.
24857: A custom solution to selectively enable/disable alerts is recommended for some use cases
Excessive dataserver memory usage can be caused when a continuously growing set of non-repeating clients is experienced. For customers that have a use case that calls for a constantly churning set of non-repeating clients, RTView Support team can apply a custom configuration that will prevent performance degradation by disabling selected alert indices. Please engage RTView support through normal channels.
24862: Regex parsing errors in soleventmodule.log
The following events may produce errors in the solaceeventmodule.log if the logging level is set to warn or lower: VPN_MQTT_RETAIN_CACHE_FAILED - can occur if the retain cache name includes special characters. This is a generic problem with most events. VPN_MQTT_RETAIN_CACHE_FAILED_CLEAR - can occur if the retain cache name includes special characters. This is a generic problem with most events. SYSTEM_ROUTING_CSPF_NBR_DOWN - can occur if the neighbor hostname includes special characters. This is a generic problem with most events. CLIENT_CLIENT_BIND_SUCCESS CLIENT_CLIENT_BIND_FAILED VPN_SOLCACHE_MSG_BUS_CONNECTED - MQTT retain is missing 'Cache Instance (1)' field VPN_SOLCACHE_STATE_CHANGE -MQTT retain is missing 'Cache Instance (1)' field
Solution Package for TIBCO BusinessWorks
17719: Problems viewing servers when using named filter with domains
The BW Named Filters mechanism allows you to specify multiple index fields (out of domains|servers|engines|processes|activities) and values on which to filter. However a domain filter cannot be used with server data. There is no Domain column in the server data table on which to filter, because a server can have multiple domains. So if you set up a filter with a domain field, e.g. <bwfilter name="domainfilter"> <domains> <domain name="domain01"/> <domain name="domain02"/> </domains> </bwfilter> ... it will work correctly on engines, process, and activities, limiting the display to engines in domain01 or domain02. But on the servers pages it will have no effect, and while it is selected, the Servers selector dropdown on all the pages will be blank.
Solution Package for TIBCO Hawk
20568: HawkAlerts cache sometimes out of sync with with alerts in the Hawk Display
There are 2 cases where the HawkAlerts cache may not be in sync with the Hawk Alerts on an agent: 1. If RTView does not receive any updates for a HawkAlert for the amount of time specified in $rtvHawkAlertClearTime + $rtvHawkAlertDeleteTime (defaults to 10 days + 1 hour), it is removed from the HawkAlerts cache. 2. When using a sender to update the HawkAlerts cache, any events that occur when the sender is running but the receiver is not will be missed. To fix this, restart the sender after any restart of the receiver.
Solution Package for VMWare
20594: MySQL historian requires NO_BACKSLASH_ESCAPES
The VMWMON package collect metrics about disk usage on Windows systems, which results in the backslash (\) character being present in the data collected. MySQL databases will not correctly store data with a backslash (\) unless the NO_BACKSLASH_ESCAPES mode is enabled.
|
|