Hardware Sentry KM for PATROL

Release Notes for v1.7.00

Home  Previous  Next

What's New

New Features

LUN Monitoring: A new class dedicated to monitoring LUNs has been added. A LUN is a volume on a disk array that is mounted by the HBA card and seen as a  local disk by the operating system. This new class reports on the availability as well as the number of available SAN paths to the volume.
Deactivate Device Class: This new KM Command enables administrator to discard one or several device classes from  monitoring. In that case, health metrics for these devices will not be taken into account by Hardware Sentry KM and thus, no alarm will be triggered upon a device failure.
Full Intelligent Platform Management Interface (IPMI) support: The KM is now capable of monitoring various servers, in-band and out-of-band, through native IPMI (typically newest IBM xSeries servers, Sun blade chassis, Cisco UCS servers, etc.).
Full support for BMC ProactiveNet Performance Management (BPPM 8.5): Hardware Sentry KM now fully and natively integrates with the latest iteration of BMC's monitoring framework. It includes the required metadata that described the classes and parameters for accurate interpretation, correlation and trending in ProactiveNet.
Missing device alert after n-times: Users can set the number of times a threshold has to be breached in a row before an alert is triggered on a missing device. This is used to prevent false alerts on missing devices.
Local host monitoring: The monitoring of the local host can now be removed.


Cisco UCS: Exhaustive monitoring of Cisco UCS chassis, Cisco UCS Interconnect Switch, Cisco UCS B-Series blade servers, and Cisco UCS C-Series rackmount servers. The KM leverages the native Cisco UCS XML API for the Interconnect Switch and IPMI for the servers.
Hitachi BladeSymphony: The product monitors Hitachi BladeSymphony servers. It reports on fans, power supplies, temperatures, voltages, blade and partition overall status.
IBM VIO Server: The KM is now able to monitor the VIO Server in IBM AIX systems without requiring access to an unrestricted shell.
Disks in VMware ESX4 servers: The KM supports the monitoring of VMware ESX4 servers and the monitoring of physical disks that are only listed in manufacturer specific classes.
IBM Director 6.2: IBM xSeries servers running Windows or Linux with the IBM Director Platform Agent 6.2 are fully covered.
Dell EqualLogic PS Series: The product now supports the monitoring of Dell EqualLogic PS Series disk arrays
HP SuperDome: The product now provides the monitoring of power supplies and fans. An overall powering and fan status is shown, along with individual component status, where available.
HBAs and LUNs on Windows systems: Full monitoring of Host Bus Adapters  (status, link status, link speed, traffic, and LUNs) on all Windows systems through WMI (an SMI-S Provider is no longer required)
HBAs and LUNs on Linux systems: Full monitoring of Host Bus Adapters (status, link status, link speed, traffic, and LUNs) on all Linux systems through Emulex and QLogic command line tools.

Changes and Improvements

Brocade SAN Switches: The KM now collects information from Brocade Specific MIBs instead of the industry standard Fibre Alliance MIB. This allows better port identification, adds remote host information.
Cisco Switch Monitoring: The SSH client has been modified so that it can process large command outputs without timing out. Due to the large number of ports on an MDS9513 switch, this switch is particularly affected by this issue. The timeouts have also been lengthened.
Remote Hosts Monitoring Limitation: The number of monitored remote hosts is now unlimited.
LUN monitoring: LUNs are no longer represented by the "Logical Disk" (MS_HW_LOGICALDISK) class but with a dedicated  "LUN" (MS_HW_LUN) class.
Improved discovery performance: The discovery processes for the managed elements run concurrently to improve the responsiveness of the KM.
Manual trigger of a discovery: The KM Command "Trigger a KM Discovery" has been replaced by a new one named "Trigger a Discovery" that forces a discovery for the selected host only.
Configurable timeouts: Discovery and collection timeouts are now configurable through configuration variables.
Large disk arrays: The code has been optimized to better handle very large disk arrays.
Improved missing device detection: The KM no longer triggers missing device alerts for temperature and voltage sensors, logical disks, enclosures, LEDs.
Quantum Tape Libraries: Parameters are activated based on whether a valid status can be collected. Descriptive parameter names were also added (Overall Cooling Status, instead of Fan Status).
New StatusInformation parameter: Object labels no longer display the current status of the device for optimized integration with BMC ProactiveNet and BMC Portal. The device status is now reported in the StatusInformation text parameter added to all classes.

Fixed Issues

Sun Solaris Systems: LUNs hosted by a EuroLogic disk array were mistakenly discovered as internal physical disks of Sun Solaris systems. As the status of external SAN disks cannot be determined by the server (the previously collected status was meaningless), EUROLOGC Disks are now excluded from all monitoring.
HP Proliant Servers: Memory modules were not properly monitored due to an issue with version 8.4 of the HP Insight Management Agent for Windows. For all versions, from 8.25 or later, of the HP Insight Management Agent, the monitoring of the memory modules has been improved to provide more information about each module as well as a more detailed status.
IBM EXP3000 External Bays (Linux): Disks in an external IBM EXP 3000 drive enclosure (directly attached to the server) were not taken into account in the monitoring of Linux servers. This was caused by the IBM Director Agent with ServeRaid extensions which failed to report on the disks connected to ServeRAID controller but located in an external drive enclosure.
The product now uses MegaCli instead of the IBM Director Agent to collect information about the physical disks located inside the IBM EXP 3000 drive enclosure.
VMware ESX Servers: Non-Fan sensors of type "Percentage" were being monitored as if they were fans (i.e. System Management Module Error Percentage) and thus creating false alerts.
Cisco SAN Switches: The product now properly detects the MDS 9509 chassis enclosure.
NetApp Filers: The size of Logical volumes greater than 2TB were not displayed correctly.
IBM AIX: Overall memory error count (Memory Controller) was not collecting correctly.The following error message would appear in the System Output Window in the PATROL Console: "No collect value available for /MS_HW_MEMORY/MS_HW_IBMAIXhdfllocalhost_MemoryController".
Invalid username/password error messages and warning messages in the System Output Window of the PATROL Console: several issues were fixed around a problem with versions 3.7.10, 3.7.20 and 3.7.30 of the PATROL Agent that are affected by a bug in the management of the credentials used in various code execution scenarios:
1.The KM failed to properly identify the versions of the PATROL Agent that are affected by this bug. Because of this, the KM did not use the workaround code path appropriately.
2.One of the workarounds used by the KM to circumvent this problem with the PATROL Agent was to force the impersonation as the default account of the PATROL Agent. But this workaround generated other "invalid username/password" error messages when the PATROL Agent's default account was specifically configured to prevent it from login in. A new configuration variable has been introduced to disable this workaround in this very specific case.
3.Also, an unnecessary warning message about a missing VBS file was displayed during the discovery process on UNIX, Linux and OpenVMS systems.
IBM Director 5.20 (Linux): The IBM Director Agent sometimes returned a temperature threshold of zero degrees. This would cause the thresholds to be incorrectly set. Hardware Sentry KM for PATROL now ignores temperature thresholds of zero.
Thermal Margin Sensors in Intel servers: The KM now ignores all negative temperature values. Temperature monitoring is now performed exclusively by the temperature status for these temperature sensors.
HBA cards using the SMI-S Compliant HBA Connector:
Qlogic SMI-S providers reported the status of the HBA FC Ports as "Error" when the link status was down. This specific error condition is now ignored to prevent false positives.
Emulex SMI-S providers do not set the link speed to zero if a link is down and thus link failures were not being reported. The connector will now report a link down if the status is link down or if the speed is zero.
The QLogic SMI-S provider included with the IBM Director Agent for Linux and the Emulex provider included with VMware ESX 4 did not include all the classes usually used to identify HBA Cards. The detection criteria have been adjusted for these providers.
The connector was not activated and/or FC ports were not discovered on some systems if the link was down at time of discovery. The connector is now activated and ports discovered even if link is down.
Manufacturer and Model information were not displayed for some Emulex HBA cards.
IBM Director 5.20 (Windows): Detection criteria have been fixed to correctly identify IBM Director 5.20 systems.
Sun Mx0000 (XSCF) Systems: Fixes location identifiers for voltage sensors so that all sensors and their correct location are shown; and adds temperature monitoring with default thresholds (as no thresholds are provided by the XSCF) of:
Ambient Temperature - Warning: 40 °C - Alarm: 45 °C
CMUs - Warning: 55 °C- Alarm: 60 °C
CPU Cores - Warning: 85 °C - Alarm: 100 °C
Remote MPIO disks reported as local on Windows. Any MPIO disk (LUN Multi-Path Disk Device) is now excluded from the monitoring process.
Power Supplies in McData Fiber Switches are now properly monitored.
The HP "HSV"-labeled disks are now excluded from all local monitoring as remote LUNs on HP EVA disk arrays should not be considered as local disks. Sentry Software provides a specific monitoring solution to properly monitor HP EVA disk arrays: HP EVA KM for PATROL
A device ID number (DeviceID) is now shown for missing objects. It is displayed in the device infobox as well as in the device health report
DegreesBelowWarning is now activated if custom thresholds are set on a temperature parameter.
The tentative detection of the SNMP Community is now performed after the startup delay (if any is configured).
VMWare ESX servers: Fixed duplicated instance problems.
The collection of link status, speed and duplex mode from dladm is now enabled, if not available from kstat or ndd
Fans and power supplies shelves that have no associated power supply or fan module are excluded.  This prevents the Hardware Sentry from discovering invalid fans and power supplies, e.g. an "embedded access" fan / power supply.  These invalid instances also tend to go missing and generate alerts.
Power Supply Status collection fixed.