Spectrum Hotfix patch H01 for 9.3 - fixes and enhancements - Part 2

Document ID : KB000018970
Last Modified Date : 14/02/2018
Show Technical Document Details

Back to Part 1 of the document.

Fixes to Login User Authentication

Symptom: When CA Spectrum is integrated with Windows 2008 R2 LDAP server, non-LDAP users are unable to log in even if "Allow User to Log in if No LDAP User found" setting is set to "Yes".
Resolution: Enhanced CA Spectrum to support integration with Windows 2008 R2 LDAP server. Non-LDAP users can now log in to CA Spectrum.
(H01, 256543, 21365347-1)

Fixes to Modeling

Symptom: The interface Name was not properly set while doing reconfigure.
Resolution: Implemented changes to set the interface name properly while doing reconfigure.
(H01, 213494, 21271790-1)

Symptom: When a WA_link model is connected to a high speed interface, bandwidth values are not displayed correctly.
Resolution: WA_link model connected to a high speed interface will display bandwidth values appropriately.
(H01, 216307, 21229554-1)

Symptom: When multiple ports have the same MAC addresses, Spectrum does not create any connections.
Resolution: If there is an Ethernet port among those remote ports which have same MAC Addresses, give preference to that Ethernet port and create connection with that port.
(H01, 254357, 21516092-1)

Fixes to Modeling Gateway

Symptom: AutoDiscovery Configurations which has SNMPv3 profiles defined in the community string with escape characters like &, <,>, ", ' are not encoded when exported.
Resolution: AutoDiscovery Configurations which has SNMPv3 profiles defined in the community string with escape characters like &, <,>, ", ' are now encoded when exported.
(H01, 256153, 21493107-1)
(H01, 272827)

Fixes to Multicast Manager

Symptom: The SpectroSERVER is processing notifications on unchanged attributes for Multicast Manager.
Resolution: The SpectroSERVER does not process notifications on unchanged attributes for Multicast Manager.
(H01, 225976)

Fixes to Navigation Hierarchy

Symptom: The Global Collection hierarchy may not display due to an exception found in client log.
Resolution: The Global Collection hierarchy was changed to single thread construction.
(H01, 225404, 21384872-1)

Fixes to Network Configuration Management

Symptom: A few devices are moving out of the custom device families after running "reevaluate NCM device family".
Resolution: If a device is placed in a custom device family, then it will stay in same family even after running "reevaluate NCM device family".
(H01, 216294, 21280082-1)

Symptom: NCM upload is failing on Cisco 837 devices with IOS 12.3.
Resolution: Improved Spectrum's handling of 12.3 devices.
(H01, 219799, 21288190-1)

Symptom: CPU usage by the OneClick process always remains high because of large NCM capture files.
Resolution: Introduced the pollRate attribute which can be set to a higher interval for larger configuration files.
(H01, 220371, 21195063-1)

Symptom: EOF errors are sometimes seen in $SPECROOT/NCM/NCMSERV.OUT when capturing a NCM configuration.
Resolution: Changes were made to allow the buffer time to process before reading begins. This requires a new file to be created in $SPECROOT/NCM/ncmprops with the following 3 settings:com.mindbright.sshcommon.SSHSCP1.readFully.initstat1=truecom.mindbright.sshcommon.SSHSCP1.readFully.initwait=10000com.mindbright.ssh2.SSH2InternalChannel.usejavaapi=true.
(H01, 251542, 21508034-1)

Symptom: The SpectroSERVER is crashing in NCM firmware upload task when the flag "NCM_SnmpFirmwareUploadDebugFlag" is set to true.
Resolution: The NCM firmware upload task runs successfully with "NCM_SnmpFirmwareUploadDebugFlag" set to true.
(H01, 254554, 21482928-1)

Symptom: An NCM file transfer timeout error is reported in the OneClick console, if the device firmware upload time exceeds the hard coded value of 66.66 minutes.
Resolution: The NCM firmware upload timeout value was made configurable in Spectrum so that the users will now be able to modify or set appropriate timeout values from the OneClick console.
(H01, 262574)

Symptom: When NCM compares the running versus previous configurations for Cisco IOS and SSH device families, NCM is incorrectly considering the certificate or password 7 changes as valid and highlights them in the NCM configuration differences window.
Resolution: NCM displays the certificate or password 7 changes appropriately under the "Irrelevant" section in the configuration differences window.
(H01, 264206, 21315816-1)

Fixes to Normalized Resources

Symptom: When the SysEdge flag is set to 0x4000 the file/dir size has values that should be shown in GB are shown in OneClick as MB.
Resolution: When the SysEdge flag is set to 0x4000 the file/dir size having values in GB will be shown correctly.
(H01, 225624, 21502515-1)
(H01, 225389, 21502515-01)

Fixes to OneClick

Symptom: The OneClick console could become unresponsive due to a rare deadlock scenario.
Resolution: Resolved a possible deadlock scenario caused by an incorrect locking order while managing relations of access groups and privilege roles.
(H01, 116208, 21089658-1)

Symptom: OneClick Client inactivity timeout functionality is broken and the users are not dropped at the OneClick Server side thereby leading to unnecessary network traffic.
Resolution: OneClick Client inactivity timeout functionality will work appropriately and the users are dropped at the OneClick Server side.
(H01, 204452, 21380344-1)

Symptom: Unprivileged users can see chassis modules in some scenarios, thereby resulting in "Unable to render topology" error.
Resolution: Unprivileged users will not be able to view the chassis modules in any situation and so there will not be any topology rendering errors.
(H01, 214478, 21438124-1)
(H01, 223279, 21386665-1)

Symptom: In the Client Details and Client Log windows for OneClick, when using a Load Balancer, the HostName field reflects the name of the Load Balancer, NOT the name of the actual client that is logged in.
Resolution: The actual IP address is retrieved from X-Forwarded-For header field and given to NameLookUpThread.
(H01, 215911, 21393032-1)

Symptom: Null Pointer Exceptions related to the progress bar dialog are randomly reported in the OneClick debug console.
Resolution: Random Null Pointer Exceptions related to the progress bar dialog are resolved and so the OneClick debug console logs will not have these exceptions reported.
(H01, 217201, 21194072-1)

Symptom: General tomcat performance failure to see alarm updates.
Resolution: Historical event processing has been placed onto its own thread pool for processing; the global thread pool has also been increased to ensure robust processing of SpectroSERVER updates.
(H01, 219807, 21341680-1)

Symptom: In the hierarchy, top level nodes such as Configuration Manager lack the plus icon to expand. An IllegalArgumentException is seen in client log.
Resolution: The IllegalArgumentException is now being handled.
(H01, 219811, 21389177-1)

Symptom: The 'Schedule Maintenance...' dropdown menu option is unavailable for users who have the 'Model Maintenance Mode Schedule' privilege, but don't have the 'Model Maintenance Mode' privilege.
Resolution: The 'Schedule Maintenance...' dropdown menu option is now solely dependent on having the 'Model Maintenance Mode Schedule' privilege.
(H01, 224475, 21480327-1)

Symptom: User groups do not remain expanded after updates.
Resolution: User groups now remain expanded after updates.
(H01, 225405, 21328225-1)

Symptom: In large Distributed SpectroSERVER configurations, tomcat could run out of memory when a client opens Event Configuration Editor for the first time.
Resolution: Limited the number of simultaneous domains in which events are gathered.
(H01, 225407)

Symptom: OneClick server goes out of memory due to caching of JDBC4PreparedStatement objects in memory.
Resolution: Modified PoolPreparedStatement setting, defaulted to false for avoiding caching of JDBC4PreparedStatement objects in memory.
(H01, 227181, 21228378-1)

Symptom: Tomcat process can abort with a non-detailed JVM crash file.
Resolution: Only seen when using earlier JVM version, upgrade Spectrum which includes updated JVM and enhanced debugging.
(H01, 251924, 21276235-1)

Symptom: Very rare, intermittent and temporary OneClick red-boxing of landscapes. Resolution: Isolated and resolved a very specific scenario in which a single client request could delay other requests from being processed. (H01, 281953, 21538402-1)

Fixes to Performance Manager (SPM)

Symptom: Tests are removed from the testTable while updating the list at the time of switching. When the SpectroSERVER is switched, some tests will not start running until those are removed from the testTable.
Resolution: Tests will not be removed from the testTable at the time of switching.
(H01, 213482, 21402274-1)

Symptom: When a user creates a SPM HTTPS test from Spectrum, the username and password fields are marked with an '*' indicating they are required fields but users can simply not enter anything in the fields and create the test as desired.
Resolution: User Interface validation for username and password fields is now made mandatory for HTTPS test creation.
(H01, 215892, 21376024-1)

Fixes to Policy Manager

Symptom: In Policy Manager settings, when the attribute ifModelNameOption/ifModelNameOptionSecondary is selected, the corresponding value field attribute is shown in decimal instead of Hex which causes policy enforcement to have no effect.
Resolution: The attribute value is now converted to Hex and can be used to enforce policy on the devices without any problem.
(H01, 221957, 21468936-1)

Fixes to QoS Manager

Symptom: Spectrum QoS Discovery is not modeling QoS Traffic Classes and Behavior models when the MatchStatement MIB value is not present in the device.
Resolution: Spectrum QoS Discovery will now model QoS Traffic Classes and Behavior models.
(H01, 227973, 21428396-1)

Fixes to Report Manager

Symptom: In Report Manager, the Shared Media Link model type is considered a device, even though it does not fall into the Device category.
Resolution: In Report Manager, the Shared Media Link model type is not considered as a Device type.
(H01, 221932, 21267451-1)

Symptom: When exporting the Availability Selected Devices and Availability Selected Models report to Excel Data only or CSV format, the header and actual data are not in line with each other.
Resolution: When exporting the Availability Selected Devices and Availability Selected Models report to Excel Data only or CSV format, the header and actual data will be in line with each other.
(H01, 226440, 21498145-1)

Symptom: When running the Service summary by service customer report, the results shown are incorrectly.
Resolution: When running the Service summary by service customer report will show the correct results.
(H01, 226866, 21317573-1)

Symptom: In some alarm reports, report header is not rendered when the query returns only one record.
Resolution: Report changes are done such that the header is shown when query returns one or more records.
(H01, 227954)

Symptom: In Report Manager, Service Availability reports are not exporting into CSV in the proper format.
Resolution: Report Manager Service Availability reports are properly formatted in CSV output.
(H01, 255451, 21466540-1)

Symptom: The Business Objects Administrator password is in readable format in HTML source of Business Objects Integration page.
Resolution: The Business Objects Administrator Password is hidden and cannot be seen in HTML source of Business Objects Integration page.
(H01, 256150, 21352796-01)

Symptom: Some OneClick users cannot access CA Spectrum reports because of missing CA Report Viewer group membership in CA Business Intelligence.
Resolution: Report Manager will ensure that all OneClick users are member of CA Report Viewer group in CA Business Intelligence.
(H01, 256152)

Symptom: On a low disk space machine, the Spectrum installer warns of wrong disk space requirements for CABI.
Resolution: The Spectrum Installer updated to reflect correct disk space requirements.
(H01, 256154)

Symptom: Report Manager has duplicate monitoring options for Business Objects Performance.
Resolution: Removed an unrelated option from Business Objects Integration to resolve the duplicate monitoring options issue.
(H01, 258041)

Symptom: In Report Manager, bucket processing is very slow.
Resolution: Resolved an issue that affected the database table setup to speed up alarm bucket processing.
(H01, 258953, 21135691-02)

Symptom: The Tomcat server is a taking long time to complete loading for report manager.
Resolution: The Spectrum report manager is enhanced to finish starting up faster.
(H01, 271174, 21375381-02)

Fixes to SNMPv3

Symptom: When the Authentication Password and Privacy Password are modified in a SNMPv3 profile template and a trap is sent with changed passwords, that trap is not displayed as an event in the SpectroSERVER.
Resolution: After modification of Authentication Password and Privacy Password, an event is displayed when a trap is sent with the new Authentication Password and Privacy Password.
(H01, 213491, 21305581-2)

Fixes to Service Manager

Symptom: Rule was not set properly because the service condition was not calculated properly.
Resolution: Rule is now set correctly and the service condition is also calculated properly.
(H01, 213488, 21260939-1)

Symptom: Cannot find some Service outages that have short duration times (around 5 seconds) in the OneClick Console, Outage History, Recent Outages.
Resolution: The Outage History table had a filtering bug. The 5-second outages are now displayed even when the filter is selected.
(H01, 269782, 21271848-1)

Fixes to SpectroSERVER

Symptom: Intermittent Archive Manager deadlock and crash, preventing events from being stored in the MySql database.
Resolution: Resolved possible deadlock processing events from the SpectroSERVER and enabled queuing to improve SpectroSERVER connection stability. Changed to using dynamic thread-safe MySql libraries on Linux and Solaris.
(H01, 208981, 21370396-3)
(H01, 258046)

Symptom: SpectroSERVER may crash during NCM capture.
Resolution: SpectroSERVER doesn't crash during NCM capture.
(H01, 214065, 21415741)

Symptom: 1. Intermittent SpectroSERVER crash when updating trap director cache due upon connection establishment.2. SpectroSERVER memory growth due to trap director forward attempts to unconnected landscapes.3. Secondary SpectroSERVER memory growth due to combination of hot-standby and trap director.
Resolution: 1. Resolved crash when updating trap director cache.2. Resolved several reasons for memory growth when trap director is enabled.3. Disabled trap forwarding on hot-standby secondary SpectroSERVER.
(H01, 214104, 21397144-1)

Symptom: If the device is in maintenance mode and it is rebooted, device reboot event is generated when the device is taken out of maintenance.
Resolution: If the device is in maintenance mode and it is rebooted, device reboot event is not generated when the device is taken out of maintenance.
(H01, 218888, 21469269-1)

Symptom: SpectroSERVER may crash while creating an HTTP RTM_Test that has an incorrect URL entry (just slash - "/").
Resolution: SpectroSERVER now handles an incorrect URL entry.
(H01, 220866, 21460587-1)
(H01, 221457, 21460587-1)

Symptom: VLAN discoveries can take a very long time to complete, holding up threads needed for other scheduled tasks.
Resolution: Resolved issues where VLAN discoveries can become intertwined, freeing threads needed for other scheduled tasks.
(H01, 221847, 21422039-1)

Symptom: SpectroSERVER is crashing during the discovery of VHM LPAR devices.
Resolution: Discovery of VHM LPAR devices will not crash the SpectroSERVER.
(H01, 225103, 21324639-1)

Symptom: SpectroSERVER crashes during startup if the Secure Domain Connector list is empty in VHM.
Resolution: SpectroSERVER will not crash during startup when the Secure Domain Connector list is empty.
(H01, 225106, 21397954-1)

Symptom: Secure Domain intermittent instability, due to rare problems on both SpectroSERVER- and SDC-sides of the connection.
Resolution: Numerous improvements to Secure Domain architecture, including: 1. Improved resiliency of the secure tunnel.2. Changed SDM threading architecture to significantly reduce SpectroSERVER thread usage.3. Changed to latest version (4.3.2) of CAPKI secure socket library.
(H01, 225658, 21280772-1)

Symptom: The SpectroSERVER was crashing occasionally in an identified area of Connection Management code.
Resolution: The SpectroSERVER no longer crashes in the area of Connection Management code where a problem has been identified and fixed.
(H01, 226522, 21431554-1)

Symptom: The connection between the Central Spectrum Server and the Remote Operations Servers was getting lost frequently. Resolution: A new configuration (ros_heartbeat_timeout) has been added in .vnmrc file, which specifies the maximum number of seconds the Remote Operations Manager will wait for a heartbeat from the Remote Operations Server, before a contact lost alarm is raised. (H01, 252049, 21395028-1)

Symptom: When a maintenance window is scheduled for device with a 24 hour duration and for a future start date, on that future date, the device is not going into maintenance mode.
Resolution: When a maintenance window is scheduled for device with a 24 hour duration and for a future start date, on that future date the device will be put into maintenance mode.
(H01, 252058, 21368331-1)

Symptom: Spectrum polls devices when taken out of maintenance and the polling status is set to off.
Resolution: Spectrum does not poll the device when it is taken out of maintenance and the polling status is set to off.
(H01, 252084, 21421255-1)

Symptom: A huge increase in memory used by the SpectroSERVER process is observed upon some discoveries, if there are Real-Time Global Collections with many AND/OR in the search Criteria.
Resolution: A cause of potential huge increase in memory used by the SpectroSERVER process upon some discoveries, if there are Real-Time Global Collections with many AND/OR in the search Criteria, has been identified and fixed.
(H01, 252090, 21522535-1)

Symptom: The "Create schedule" dialog displays a time zone, which is confusing, given that the time specified will be, for each SpectroSERVER, for its local time zone.
Resolution: The "Create schedule" dialog no longer displays a time zone (time specified will be, for each SpectroSERVER, for its local time zone).
(H01, 255705, 21538851-1)

Symptom: Security strings containing ADMIN are not retained after model type conversion using Newmm.pl.
Resolution: Security strings containing ADMIN are now retained after model type conversion using Newmm.pl.
(H01, 259284, 21510240-1)

Symptom: The SpectroSERVER crashes when activating certain Nortel BayStack devices. (This has been observed after multiple SpectroSERVER database migrations.).
Resolution: The SpectroSERVER no longer crashes when activating these Nortel BayStack devices. If the exact scenario is hit (which was observed after multiple SpectroSERVER database migrations), a message is reported in VNM.OUT.
(H01, 260200, 21495269-1)

Symptom: The SpectroSERVER takes a long time to activate models during large network outage.
Resolution: Resolved the activation issue which was caused by external reads during Port IPLS attribute calculations and added additional logging when running into thread exhaustion scenarios.
(H01, 262621, 21525389-1)

Symptom: Intermittent SpectroSERVER crash when computing alarm impact.
Resolution: Improved handling of alarm impact objects to prevent crash.
(H01, 264314, 21567402-1)

Symptom: Back-to-back (ie. daily, 24 hour) maintenance schedules trigger devices to come out of maintenance temporarily, introducing a small window when alarms can be generated.
Resolution: Maintenance mode intelligence now detects when back-to-back schedules exist, and devices are no longer taken out of maintenance temporarily.
(H01, 264764, 21477152-1)

Symptom: The SpectroSERVER does not start in Remote Operation Server installation on Linux or Solaris platforms because of the newly created dependency for SpectroSERVER on libsdmmgr.so.1 shared lib.
Resolution: The dependency is not mandatory for the SpectroSERVER to run, hence the dependency on libsdmmgr.so.1 has been removed.
(H01, 268862)
(H01, 270983, 21596507-1)

Symptom: SpectroSERVER may crash during SpectroWatch evaluation.
Resolution: SpectroSERVER doesn't crash during SpectroWatch evaluation.
(H01, 274699, 21596582-1)

Symptom: SpectroSERVER may crash during VPN discovery.
Resolution: Intermittent SpectroSERVER crashes during VPN discovery will not be seen.
(H01, 285429, 21621479-1)

Symptom: SpectroSERVER crashes when it is processing one of the varbinds from a malformed trap.
Resolution: SpectroSERVER does not crash when a malformed trap is processed.
(H01, 285448, 21564849-1)

Fixes to SpectroSERVER Fault Tolerance

Symptom: Event discriminators are not synchronized correctly in a fault tolerant environment.
Resolution: Event discriminators will be synchronized correctly in a fault tolerant environment.
(H01, 225081, 21259279-1)

Symptom: OneClick shows SpectroSERVER as incorrectly failed over.
Resolution: Resolved rare deadlock/performance problem that could cause the SpectroSERVER to stop processing CORBA requests.
(H01, 264451, 21550062-1)

Fixes to System Center Operations Manager

Symptom: SCOM Connector fails to send source machine name to Spectrum for SCOM agent watcher group alerts that are asserted on VNM model.
Resolution: SCOM Connector sends the source machine name to Spectrum for SCOM agent watcher group alerts that are asserted on VNM model.
(H01, 256071, 21364645-1)

Fixes to Topology

Symptom: ESX Host containers, the topology view zoom % is not as per user preferences.
Resolution: Topology view zoom % is as per user preferences for all container types.
(H01, 214977, 21451991-1)

Symptom: Client hierarchy can take several minutes to fully populate.
Resolution: Code making a native API method call has been replaced with a faster non-native method call.
(H01, 219802, 21355167-1)

Fixes to Trap Forwarding

Symptom: BGP BackwardTransition Event is not showing any data when varbind instance is from IP V6 address.
Resolution: Data is populated appropriately for the BGP BackwardTransition event when the varbind instance is from IP V6 address.
(H01, 220089, 21434448-1)

Symptom: Trap Director stops forwarding traps to some remote models if landscapes are removed from its list of available landscapes.
Resolution: Simplified Trap Director flow. Improved debugging information is now available. Available landscapes are now re-evaluated for every trap, making the process more resilient when there are landscape map and landscape availability changes.
(H01, 264453, 21325883-1)

Fixes to VPN Manager

Symptom: In case of overlapping VPN configuration, Spectrum is adding only one VPN name to the VPN Site name thereby causing the VPN Site names to be incorrect.
Resolution: In case of overlapping VPN configuration, Spectrum will append all the overlapping VPN names to the VPN Site. This will be effective after the existing VPNs are removed and a new VPN Discovery is made.
(H01, 227200)

Symptom: VPN Manager configured in Distributed SpectroSERVER setup is not showing the associated site names in Component Details View, Spotlight View, and the Locator Search.
Resolution: VPN Manager configured in Distributed SpectroSERVER setup will show the associated site names in Component Details View, Spotlight View, and the Locator Search.
(H01, 227201, 21282145-1)

Symptom: VPN Manager under DSS setup is showing empty Route Statistics.
Resolution: VPN Manager under DSS setup will now show the Route Statistics correctly.
(H01, 227203)