This knowledge document only lists the fixes and enhancements to pre-existing functionality that Spectrum 18.104.22.168 (aka Service Pack 3, H10) will be delivering. The full and final Software Release Notice (SRN) contains requirements, special considerations and installation instructions. Please read thefull SRN once it is available for download along with 22.214.171.124.
DETAILS OF 9.2.3 CONTENTS
Cisco Nexus Certification (9.2.3, 85823, 85825, 85827, 19113)
The Cisco Nexus Certification has been updated to include 9 additional devices including the Cisco Nexus 5548UP, Cisco Nexus MDS 9216A, Cisco Nexus MDS 9140, Cisco Nexus MDS 9120, Cisco Nexus MDS 9124e, Cisco Nexus MDS 9134, Cisco Nexus MDS 9216i, Cisco Nexus MDS 9124, and Cisco Nexus MDS IBM Blade Switch.
The device type name for Cisco Nexus 5548P, Cisco Nexus 5020 AC, Cisco Nexus 5020 DC, devices which have been certified already are modified to appropriate names as indicates in cisco website.
Cisco Nexus Firmware Upgrade (9.2.3, 122457)
Spectrum has been enhanced to support Cisco Nexus Firmware version 6.1(1). Features include updated MIB and Trap support.
Lancom 1781A-3G Certification (9.2.3, 85774)
Spectrum has been enhanced to support Lancom 1781A-3G device with the LancomLCOS model type.
HP A10508 Certification (9.2.3, 115961)
Spectrum has been enhanced to support HP A10508 device with the HpH3cDev model type.
Digi TransPort WR21 Certification (9.2.3, 103308)
Spectrum has been enhanced to support Digi TransPort WR21 device with the DigiTransPort model type.
Radware LinkProof 1008 Certification (9.2.3, 129371)
Spectrum has been enhanced to support Radware LinkProof 1008 device with the Radware_LP model type.
Cisco ACE Certification (9.2.3, 30991)
Spectrum has been enhanced to support 4 Cisco ACE devices including the Cisco ACE 10K9, Cisco ACE 20K9, Cisco ACE 30K9, Cisco ACE 4710K9 with the CiscoACE model type.
Dell OpenManage 7.1 (9.2.3, 128965)
Spectrum has been enhanced to support Dell OpenManage version 7.1. Features include updated MIB and Trap support.
Adtran Device (9.2.3, 104923)
Spectrum has been enhanced with a simple certification for the Adtran Device.
Convedia RadiSys Media Server v3.1 (9.2.3, 94711)
Spectrum has been enhanced with a simple certification for the Convedia RadiSys Media Server.
Expand 6950 Device (9.2.3, 104919)
Spectrum has been enhanced with a simple certification for the Expand 6950 Device.
UnAcknowledge_On_New_Occurance (9.2.3, 120371, 21088211-1)
Added the optional ability to have acknowledged alarms be unacknowledged when a new occurence happens. This can be controlled by a new AlarmManagement model attribute UnAcknowledge_On_New_Occurence( 0x13208), which can be set via the AlarmManagement subview of the VNM model Information view, under the "Auto UnAcknowledge on New alarm Occurance". The default is FALSE, preserving existing behavior.
New Support for Microsoft MSCS and IBM PowerHA Cluster Server Management (9.2.3)
Support for Microsoft Cluster Service and IBM PowerHA clustering technology is now supported. A new distributed level node in the OneClick Explorer panel labeled "Cluster Manager" provides a central place to manage all cluster components. Cluster nodes are identified and visualization of resource groups and resources are provided via the OneClick hierarchy as well as the new cluster container located in the Topology. The Cluster Manager leverages the MSCS and HACMP cluster AIM solutions which provide the monitoring capabilities for the cluster environments.
SCOM Connector Compatible to SCOM Manager 2012 (9.2.3, 112457)
Existing functionality of Spectrum-SCOM integration, when Spectrum SCOM Connector is configured for SCOM Manager 2012, will work on the same lines as Spectrum-SCOM integration when connected to Microsoft SCOM Manager 2007. New functionality introduced in Microsoft SCOM Manager 2012 is not supported in this version of Spectrum.
CA eHealth Integration (9.2.3, 110245)
Enhanced eHealth mapping to map Interface models queried from nsIfTable which are proprietary to NetScreen devices.
Dynamic Alarm Titles for SystemEDGE traps (9.2.3, 155223)
The titles for alarms generated from following SystemEdge Traps have been enhanced to dynamically display specific information related to the cause of the trap derived from the trap specific varbinds, so that appropriate action is taken against each independent alarm:
Fixes to Alarm Management
- Symptom: Spectrum may generate an alarm on a port interface of type Atmvclnk having the "GeneratePortStatusAlarms" attribute set to "No".
Resolution: Spectrum does not generate an alarm on a port interface of type Atmvclnk having the "GeneratePortStatusAlarms" attribute set to "No".
(9.2.3, 120082, 21116550-1)
- Symptom: When a trap is received for Log File Monitor with severity Warning from SystemEDGE host, a Minor alarm is raised in Spectrum on SystemEDGE Host model.
Resolution: When a trap is received for Log File Monitor with severity Warning from SystemEDGE host, an event is created in Spectrum on SystemEDGE Host model instead of Minor alarm.
Fixes to Alarm Notifier and SANM
AlarmNotifier does not show event information when the Archive Manager is unavailable.
Resolution: AlarmNotifier will display the originating event and the clearing event, if available, when the Archive Manager is unavailable.
(9.2.3, 131210, 21129520-3)
Fixes to AutoDiscovery
- Symptom: Auto export of model discovery results does not contain the SNMP community string.
Resolution: The SNMP community string has been added to the exported discovered model results set.
(9.2.3, 141586, 21164474-1)
Fixes to CA Performance Center (NPC)
- Symptom: CA Performance Center always shows wrong version of Spectrum as 9.2.1.
Resolution: CA Performance Center will show correct version of Spectrum.
- Symptom: Spectrum's launch back URL for Performance view always uses Performance Center' s "Device" view even if Spectrum's model under context is of type "Sever" or "Workstation-Server".
Resolution: Spectrum's launch back URL for Performance view will launch "Server" view of Performance Center if Spectrum's model under context is of type "Server" or "Workstation-Server".
(9.2.3, 92643, 20953977-1)
Fixes to Cisco
- Symptom: Some UCS components may be deleted in Spectrum when the UCS AIM is restarted, resulting in possible loss of model customization and configuration.
Resolution: As a resolution to this problem, the UCS AIM was updated to include a status object which indicates if the manager is ready or not. Spectrum will process the MIB once the manager is ready, and will not delete UCS components if the manager is not in a ready state.
(9.2.3, 94706, 20603652-4)
Fixes to Condition Correlation
- Symptom: Interface condition remains critical and bad link alarm is not cleared when model is coming out from maintenance mode.
Resolution: Interface condition is proper and bad link alarm is cleared, when model is coming out from maintenance mode.
(9.2.3, 107946, 21057781-01)
Fixes to Event Management System
- Symptom: An EvFormat file does not exist for the DDM Table Error event (0x00010c1c).
Resolution: An EvFormat file for the DDM Table Error event (0x00010c1c) is added.
(9.2.3, 106695, 21059337-1)
- Symptom: Event format files of WA_SEG_CONN_NORM(0x10d57) and WA_SEG_INVALID_ASSOCIATION(0x10d58) events are not in proper format.
Resolution: Event format files of WA_SEG_CONN_NORM(0x10d57) and WA_SEG_INVALID_ASSOCIATION(0x10d58) events are in proper format.
- Symptom: Varbind mapping is incorrect for ntEventMonMatchTrap.
Resolution: Varbinds are now mapped correctly for ntEventMonMatchTrap.
Fixes to Fault Isolation
- Symptom: When two separate fault domains are merged (i.e. by adding a new connection between devices), issues are seen with the root causes and impacted devices for unresolved faults during Fault Isolation.
Resolution: Fault Isolation was enhanced to cover these cases correctly.
(9.2.3, 140345, 20948244-1)
Fixes to Install
- Symptom: The SpectroSERVER fails to start, due to missing libraries, after being updated via 9.2.0 with a newer key and re-applying the latest hotfix. A library was obsoleted in 9.2.0, and then re-introduced in a hotfix.
Resolution: Install has been updated to ensure that the obsoleted library is installed when the hotfix is re-applied.
Fixes to Modeling
- Symptom: The SpectroSERVER may not delete a stale interface during the interface reconfiguration even if the attribute HasStaleInterfaces is set to FALSE.
Resolution: The SpectroSERVER deletes a stale interface during the interface reconfiguration even if the attribute HasStaleInterfaces is set to FALSE.
(9.2.3, 124186, 21098416-1)
Fixes to Network Configuration Management
- Symptom: NCM global sync task (GST) could be in inconsistent state for landscapes with 0 devices.
Resolution: Ensured proper event, start time setup for starting of GST, as well as cleanup and stop of the GST.
(9.2.3, 103069, 20694488-1)
- Symptom: Cisco devices with IOS version format XX.X(X.XX) are not getting added under the correct device family in Network Configuration Manager (NCM).
Resolution: Cisco devices with IOS version format XX.X(X.XX) will get added under the correct device family in Network Configuration Manager (NCM).
(9.2.3, 120367, 20993522-1)
Fixes to OneClick
- Symptom: Clients with large hierarchy trees run the potential to exhaust client memory limits.
Resolution: Array space consolidation has been found in tree node creation.
(9.2.3, 105911, 21057469-1)
- Symptom: Locater tab searches in OneClick could fail if the client is running with the Java 7 runtime.
Resolution: Handled the API change to the Java 7 runtime
- Symptom: Event retrieval is slow from the OneClick event tab.
Resolution: Improved performance so that event retrieval is no longer slow from the OneClick event tab.
(9.2.3, 126234, 21140547-1)
- Symptom: Degree Sign for Cisco devices is garbled in Linux operating system.
Resolution: Degree Sign for Cisco devices is replaced by appropriate phrase to support all operating systems.
(9.2.3, 127944, 21118168-1)
- Symptom: The Performance tab is grayed out for CiscoCSS Models.
Resolution: The Performance tab is enabled for CiscoCSS Models.
(9.2.3, 129210, 21064405)
- Symptom: Alarm table on the webserver side appears to be hung, no more new alarms or notifies for old alarms (updates/clears) are seen. When multiple webservers are set up, it may happen on a single one, or on multiple ones, so that individual alarm views may show different alarm information.
Resolution: Alarm watch registration was improved to handle SpectroSERVER-OC webserver communication issues better, ensuring that alarm watches are always properly set up.
(9.2.3, 140573, 21157767-1)
(9.2.3, 143530, 21164197-1)
- Symptom: This work item added alarm table and alarm client debug capabilities, which may provide helpful information for any alarm table and alarm notification issues.
Resolution: Alarm table and alarm client debug added.
(9.2.3, 140576, 21029668-1)
Fixes to Performance Manager (SPM)
- Symptom: After a SpectroSERVER restart some traceroute tests are being duplicated.
Resolution: We now check the OnDemandTest_mh value before writing to the DB. If the value is 0x0, Spectrum will not update TesterMH attribute. If the value is not null, spectrum will update TesterMH attribute. This will prevent the duplication of the traceroute tests.
(9.2.3, 37768, 20527079-1)
(9.2.3, 122855, 21101315-1)
Fixes to Processd
- Symptom: Running the 'processd --stop' command may not stop the mysqld process when upgraded from 9.2.0 base version.
Resolution: Running ''processd --stop' command now stops the mysqld process when upgraded from 9.2.0 base version.
Fixes to Report Manager
- Symptom: If the trouble ticket ID does not contain service desk URLs, errors are thrown during alarm bucket processing. Also, sub reports will not be generated for some alarm reports where the trouble ticket id is displayed.
Resolution: Alarm buckets will get processed. Sub reports of alarm reports where trouble ticket id is displayed will be generated.
(9.2.3, 120621, 21107688-1)
(9.2.3, 123700, 21132447-1)
(9.2.3, 123703, 21117406-1)
- Symptom: Reports generated that are exported to PDF are not rendered correctly. The text in the PDF version of the report is unreadable.
Resolution: All the reports exported to PDF are now readable.
(9.2.3, 128688, 21141008-1)
- Symptom: Report Manager stops event processing when a duplicate username is detected.
Resolution: The Security handler will handle when a deleted OneClick username is recreated, therefore eliminating the issue.
(9.2.3, 129192, 21104285-1)
- Symptom: Report Manager won't initialize and throws NullPointerException at start up.
Resolution: Report Manager will handle NullPointerException during initialization.
(9.2.3, 34203, 20581147-1)
Fixes to SS and DB Tools
- Symptom: In some cases, StartSS.pl fails to start, even if the Spectrum Process Daemon (processd) is started.
Resolution: StartSS.pl no longer fails to start if the Spectrum Process Daemon (processd) is started.
(9.2.3, 144869, 21119510-1)
Fixes to SpectroSERVER
- Symptom: Memory leak in multicast code.
Resolution: Memory cleans up correctly.
- Symptom: While running Upload task from Configuration Manager the SpectroSERVER crashes.
Resolution: SpectroSERVER runs fine while running Upload task from Configuration Manager.
(9.2.3, 108804, 21055710-1)
- Symptom: When a device in Maintenance Mode is connected to a WA_Link, a situation may occur that causes a portion of the SpectroSERVER's threading to become deadlocked. It could occur when a user puts the device into Maintenance Mode or when a port connected to the WA_Link changes its status.
Resolution: Fixed the coding flaw that led to the deadlock.
(9.2.3, 126533, 21140394-1)
- Symptom: Under certain circumstances, the contents of real time Global Collections may not be accurate. In addition, real time Global Collections may cause various performance problems, such as slow model activation and high memory usage.
Resolution: Changes were done in the Real time Global Collections implementation, to resolve accuracy issues and improve performance.
- Symptom: The SpectroSERVER never raises an alarm when the device is taken out of maintenance and the disk monitoring threshold is still violated.
Resolution: The SpectroSERVER raises an alarm when the device is taken out of maintenance and the disk monitoring threshold is still violated.
(9.2.3, 142397, 21153052-1)
- Symptom: The SpectroSERVER may crash while re-evaluating the model name and if we enable the logging by setting debug_model_naming_order=TRUE in the .vnmrc file.
Resolution: The SpectroSERVER doesn't crash while re-evaluating the model name and if we enable the logging by setting debug_model_naming_order=TRUE in the .vnmrc file.
(9.2.3, 147059, 21163723-1)
Fixes to SpectroSERVER Fault Tolerance
- Symptom: When Fault Tolerant Alarm Service (FTAS) is used, the following message may erroneously be generated when the primary SpectroSERVER is starting up: Primary SpectroSERVER failed to validate the new alarms received from the secondary SpectroSERVER. The problem occurs when the primary SS has a model in it that the secondary SS does not have, and that model has an alarm on it.
Resolution: Modified the FTAS to check for this scenario and not generate the error message.
(9.2.3, 72347, 20871580-1)
Fixes to SpectroSERVER Watch Management
- Symptom: When the SpectroSERVER is restarted, any model having watches that are marked 'Active By Default' are stuck in INITIAL state.
Resolution: All watches are in Active state after a SpectroSERVER restart.
(9.2.3, 82074, 20803728-1)
(9.2.3, 102313, 21012012-01)