- Channel HP
- :
- Enterprise Business Blogs
- :
- Coffee Coaching
- :
- Coffee Coaching
- :
- HP ProLiant Gen8 Agentless Management Overview
- Subscribe to RSS Feed
- Mark as New
- Mark as Read
- Bookmark
- Subscribe
- Email to a Friend
- Printer Friendly Page
- Report Inappropriate Content
HP ProLiant Gen8 Agentless Management Overview
The Agentless Management is one of the number one new features they list about ProLiant Gen8manageability.
ProLiant servers have always had great manageability. Historically speaking, this manageability was provided by OS-based SNMP Agents or WBEM Providers and these Agents and Providers will continue to be available in Gen8.
The Nirvana of Agentless Management is that our customers will get all core management functions, for internal system components, including health monitoring AND alerting straight from iLO without ANY drivers or Agents installed. This functionality is almost effortless to configure and will function regardless of the OS or the state of the OS. All the customer needs to do is plug in the iLO network cable and plug in the power cord and the server will be inherently manageable via iLO. iLO will monitor the health of the server and send SNMP alerts.
The subsystems supported by iLO that are covered and “regardless of the operating system” are as follows:
- CPUs
- Memory
- Temperature Sensors
- Fans
- Power Supplies
- SmartArray controllers (including cache modules) and
- Both Logical and Physical Drives (internal to the server) that are connected to the SmartArray controller.
- iLO supports Inventory and Configuration, Health Status, SNMP Alerting, Pre-Failure Alerting, Redundancy Status and Alerting, and SNMP Data Collection.
All of this is done by the iLO firmware working in conjunction with the ProLiant System ROM and the SmartArray firmware. No drivers, no agents and regardless of the OS. The OS can be Windows Server 2008, a server running the Hyper-V role, a beta version of a future Window Server operating system, some “other OS” or no OS at all. The server can be in a bare-metal, pre-OS state and still have its health monitored and still have a rich set of alerting. Additionally, the server OS can even be in a halted or non-functional state
What happens when there is a system fault – like a fan failure, for example?
First, iLO will detect the fan failure. Next, iLO will log the fan failure and the loss of fan redundancy to the server’s Integrated Management Log. iLO will then send two SNMP alerts (the fan failure alert and the redundancy lost alert) to up to three alert destinations. But that’s not all. iLO will also log detailed hardware telemetry information to our new Active Health System and, if the customer has subscribed, a replacement part request can also be sent all the way back home to HP.
See the video on iLO Active Health Agents here:
How does this compare to the competition? Does anyone else have Agentless Management?
Some others may claim Agentless, but I think they are going to fall far short of what iLO provides. They may have some built-in stuff that sends a generic trap saying something like “Sensor 7 – State De-Asserted”. It’s then up to you to go figure out what is Sensor 7 and whether or not a De-Asserted state is a good thing or a bad thing. The iLO Agentless support will send the same, rich SNMP trap as the OS-based Agents. Customers will have a smooth transition from OS-based Agents to iLO Agentless.
There are still some things iLO (running as firmware below the level of the OS) doesn’t know. That’s where your Agentless Management Service comes in. Can you tell us about the AMS?
Agentless Management Service is a lightweight service that collects OS based management information and provides this information to iLO. Typically it collects server inventory information on installed applications and processes running, system name, IP address, OS installed and some NIC information that iLO has no access to. It runs in the OS context, but it is not an agent in the traditional sense. On Gen8 servers, iLO provides a significant amount of platform management information without an OS based agent. AMS will provide iLO with a richer data set that is OS specific. The information that AMS provides to iLO is a very small subset of the data that the traditional SNMP agents provided, therefore AMS is a very lightweight component. AMS consumes less memory and CPU compared to the in-band SNMP agents. Also, AMS has no network connections, therefore there are no security risks as there is no SNMP stack running in the OS context.
So how does AMS get installed and configured?
On Gen8 servers, AMS will be installed by default, in place of the OS based SNMP agents. It requires no configuration of its own. iLO configuration to enable Agentless Management will apply to AMS as well.
Is there any other functionality in AMS?
Yes, AMS will also log OS based information to Active Health System, another new feature on the Gen8 platform. It will log information like an inventory of installed applications, drivers, and services, bug check codes when there are system crashes, critical events from the OS eventlog, OS startup and shutdown events, device hot/plug events and NIC Link up/Down events. This information may be useful in root causing issues related to the platform.
For more information visit:
- Mark as Read
- Mark as New
- Bookmark
- Highlight
- Email to a Friend
- Report Inappropriate Content
I read some claims has been posted around the overall performance, including 10 per cent less energy consumption, six times faster I/O for storage and three times productivity for admins.
- Mark as Read
- Mark as New
- Bookmark
- Highlight
- Email to a Friend
- Report Inappropriate Content
iLO is definitely one of the better features on the HP Proliants. We manage a range of clients with HP servers at their premises and within our datacentre in Sydney, Australia. Being able to use iLO remote management means less time traveling and more time buidling servers, troubleshooting and fixing problems. For our clients it means we can respond faster and geography becomes irrelevant.
Mark Elliott
- Mark as Read
- Mark as New
- Bookmark
- Highlight
- Email to a Friend
- Report Inappropriate Content
When I install the 6.02 Service Pack on my DL380p G8 SLES10 SP4 system I end up with plenty of agents installed:
hp-snmp-agents
version control agent
hp-ams (agentless monitoring agent)
hp-health
Please explain how this is "Agentless"? My customers are going to ask.
By the way, my testing of sequential i/o (video streaming, high speed printing, etc.) and pdf/postscript ripping shows no difference from G6 performance.
I do not consider power usage to be a performance measurement.
- Mark as Read
- Mark as New
- Bookmark
- Highlight
- Email to a Friend
- Report Inappropriate Content
Hello sdrake.
For Gen 8 is ALL of the agents are loaded but not enabled. This means only the hp-ams service (a.k.a. hpHelper) should be activated.
What does Agentless Management mean?
- The SNMP Master has moved from the Operating System to the iLO. This provides a secure management firewall between server management operations/data and customer operations / data. The customer application is no longer subject to potential security risks that sometimes appear in OS based management stacks.
- SNMPv1 was chosen for the initial implementation due to wide acceptance of this protocol. Many customers continue to use SNMP to manage Network Switches and other devices in their infrastructure so this protocol is a natural fit. Additionally, there are many tools today that can leverage the HP supplied “Compaq MIBS”.
- The iLO is able to provide direct query and alerting via SNMP for Fans, Power, SmartArray and direct attached internal drive, Memory and Temperature. The Agentless Management Service is optional.
- The Agentless Management Service (AMS) provides OS specific detail that includes the OS Host Name (if desired for TRAPs over the iLO sysName), Operating System, Operating System Version, running processes and Network IP Addresses. Since the operating system owns the DHCP client, only the Operating System knows the IP address.
- The Agentless Management Service (AMS) leverages the mature “hp-iLO” technology for transferring data between Host (or OS) Memory and the iLO. This technology is based on industry standard PCI Express (PCIe) memory bus mastering used by other PCIe controllers such as network and storage adapters. This technology spans generations of iLO management processors and is very reliable. There is no need to enable any management protocol on the operating system if only Agentless Management via SNMP is desired.
- The iLO Agentless Management also enables a “broadcast” SNMP “Cold Start” TRAP for iLO and server resets. This feature allows Central Management Services (CMS) to quickly discover new devices or changes in devices. This technique is standard and will work for both HP System Insight Manager / HP Insight Control as well as for other SNMP based CMS applications.
- The work to embed SNMPv3 for fully encrypted sessions is almost complete and is expected to ship with iLO FW in 4CQ2012.
- For Gen 8 servers and specifically iLO 4 FW v1.10 (to be released in 3CQ2012), the iLO works very closely with the HP Platinum Plus power supplies and the HP Intelligent PDU to monitor not only server power redundancy but also data center redundancy. For the industry, HP is providing an advanced solution that allows large data center operations and administration (for Power and Cooling) to be separated from server management. This is a growing trend in Cloud Infrastructure and Computing.
Go to www.hp.com/go/ipd for more details on this technology.
- Mark as Read
- Mark as New
- Bookmark
- Highlight
- Email to a Friend
- Report Inappropriate Content
Informative post..
- Mark as Read
- Mark as New
- Bookmark
- Highlight
- Email to a Friend
- Report Inappropriate Content
This still doesn't give one so much as a clue how to use AMS.
iLO appears to only send SNMP traps. Our monitoring system does SNMP polling, so apparently I would need to find something out there that accepts SNMP traps and can act on them -- but I can find nothing that seems to be appropriate. With 100+ systems, hitting the iLO HTTPS interface and manually scanning the log every few minutes is infeasible.
Sun ILOM can be configured to send alerts via email, which is much more useful.
- Mark as Read
- Mark as New
- Bookmark
- Highlight
- Email to a Friend
- Report Inappropriate Content
Hello Anhony11, please see below a detailed answer frmo our experts. If any further comment please let us know:
"This still doesn't give one so much as a clue how to use AMS. "
The Agentless Management Service (AMS) is simply that: A small, light weight service that provides specific Operating System information such as the OS Name, Version to the iLO Management Processor. One key piece of information that the AMS provides is the Network IP Addresses of the host based NIC’s as those are maintained by the OS specific network stack.
There is nothing to “use” with AMS; This is a simple service that enriches the data that the iLO 4 management processor can provide to an application.
For more details:
Considerations for choosing HP Agentless Management
http://h20000.www2.hp.com/bc/docs/support/SupportM
HP Systems Insight Manager 7.0 and HP Agentless Management overview
http://h20000.www2.hp.com/bc/docs/support/SupportM
High Level overview of Agentless Management
"iLO appears to only send SNMP traps. Our monitoring system does SNMP polling, so apparently I would need to find something out there that accepts SNMP traps and can act on them -- but I can find nothing that seems to be appropriate."
The current version of iLO 4 FW v1.13 can alert the end application via SNMP Traps (Version 1 today, Version 3 in an upcoming release). These traps are well defined and have been in existence for multiple generations of ProLiant Servers with the legacy SNMP Management Agents installed.
"With 100+ systems, hitting the iLO HTTPS interface and manually scanning the log every few minutes is infeasible."
To simply poll the overall status of the HP ProLiant Service, you can perform a simple get on the “cpqHoMibHealthStatusArray”:
# snmpget -Os -t 1 -v1 -c <SNMP COMM STRING> <iLO IP ADDRESS> enterprises.232.11.2.10.7.0 2>/dev/null
cpqHoMibHealthStatusArray.0 = Hex-STRING: 04 04 02 02 02 02 02 04 02 04 00 02 00 00 00 02 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
This is a single “ping” of the server and gives you the overall status of the server in a single GET operation. If the first two bytes returned (e.g. 04 04 in this case) are set to “00 00”, this means the server is not powered on.
This MIB item is defined in the “cpqhost.mib” available from:
http://h18013.www1.hp.com/products/servers/managem
The definition of “cpqHoMibHealthStatusArray”:
cpqHoMibHealthStatusArray OBJECT-TYPE
SYNTAX OCTET STRING (SIZE (1..256))
ACCESS read-only
STATUS mandatory
DESCRIPTION
"The MIB Health Status Array is an array of status values representing an
overall status in element 0 follwed by server and storage status values as follows:
Octet Element Field
======== ======= =========
0 0 Aggregated Status of array elements
1 1 Status of element 1
2 2 Status of element 2
.
.
.
n n Status of element n
Status 0 - Not available
1 - Other
2 - OK
3 - Degraded
4 - Failed
.
.
Mibs are assigned an array element as follows. New items are added at the end.
0 - System Health Status (overall status as reported by who is reporting (agents or iLO))
1 - Total Aggregate (Includes IML Status)
2 - Processors (TBD 232.1.2.2.4:cpqSeCpuCondition)
3 - Memory (232.6.2.14.4:cpqHeResilientMemCondition)
4 - Cooling (232.6.2.6.4:cpqHeThermalSystemFanStatus)
5 - Sensors (232.6.2.6.3:cpqHeThermalTempStatus)
6 - Power (232.6.2.9.1:cpqHeFltTolPwrSupplyCondition)
7 - ProLiant Logs (232.6.2.11.2:cpqHeEventLogCondition)
8 - ASR (232.6.2.5.17:cpqHeAsrCondition)
9 - Drive Array (232.3.1.3:cpqDaMibCondition)
10 - SCSI (232.5.1.3:cpqScsiMibCondition)
11 - Storage Enclosures (232.8.1.3:cpqSsMibCondition)
12 - IDE (232.14.1.3:cpqIdeMibCondition)
13 - FC (232.16.1.3:cpqFcaMibCondition)
14 - Networks (232.18.1.3:cpqNicMibCondition)
15 - MP (232.9.1.3:cpqSm2MibCondition) "
::= { cpqHoSystemStatus 7 }
We hope this information is helpful,
Best, Coffee Coaching Team.
- Mark as Read
- Mark as New
- Bookmark
- Highlight
- Email to a Friend
- Report Inappropriate Content
I have a couple of DL360p Gen8 servers and I'm trying to figure out the difference of using SNMP vs AMS.
I compare two physical identical servers where one has Windows 2012 and SPP 2012.10.0 installed using hpsum.exe /use_ams and the other doesn't have any OS installed at all (it's just rebooting trying to PXE boot).
The Windows server has the "HP ProLiant Agentless Management Service" installed.
But I can't really see any additional information inside iLO of the Windows server, just the normal CPU, temp etc.
Are you saying I should see much more info in iLO once AMS is installed in the Windows 2012 OS?
- Mark as Read
- Mark as New
- Bookmark
- Highlight
- Email to a Friend
- Report Inappropriate Content
Hello Jonas,
Today, for the iLO GUI, the optional AMS Service provides the IP Address and the Network Interface Card (NIC) Status to the iLO GUI.
For the most part, the AMS Server provides a lot of legacy data such as the Operating System Name, OS Version, some PCIe Slot Information and other things that iLO cannot obtain directly from a device. This data is typically seen when using an “snmpwalk” type tool. Typically the cpqsteq.mib and cpqsinfo.mib file data is supported by AMS. The AMS also contributes information to the support of the cpqhost.mib file (such as the “Server Name”).
Below is an image that outlines exactly what you have the ability to see with Agents( SNMP agents) and Agentless Management ( iLO Firmware ) & Agentless Management with Agentless Management Service . Hoping this helps.
Coffee Coaching Team
