How to Monitor an Adaptec/Intel ICH10R RAID Controller (any SNMP notification options?)

Question

SuperMicro's Super Doctor III is the software I was looking for. It is SuperMicro's equivalent of IBM's ServerGuide or Dell's OpenManage. Unfortunately, while it supports hardware monitoring and SNMP or Email alerts it does not support running on Server Core.

---

I have a SuperMirco server that is using the MBD-X8-DTL-I-O mainboard which is equipped with one of the lovely Intel ICH10R based fake-RAID controllers (confusingly enough reported as an Adaptec device -- I'm assuming that it is some unholy marriage of the ICH10R chipset for the controller and Adaptec firmware). After some thrashing around I was able to find the correct ACHI drivers and get Windows Server 2008 R2 installed on RAID-1 setup.

Part of my confusion comes from the fact that Windows reports the controller as an "Adaptec Serial ATA HostRAID" and the Hard Drive as an "ADAPTEC RAID 1 SCSI Disk Device" whereas during the boot process the Controllers hardware clearly reports that it is actually using the Intel ICH10R ("Controller #00: HostRAID-ICH10r at PCI Bus:00...).

I don't expect much from this controller in terms of management but is there any software (3rd party or otherwise) that can monitor the health of the RAID array and notify me (preferably via SNMP to our OpenNMS server)? This server is going off-site and it would nice to know that a drive has died without physically having to look at the Blinking Lights. All I can find is the cheesy Intel RST software which I don't think has any remote monitoring and notification facilities (it also complained that it couldn't be installed due an "incompatibility error" which is strange considering that the controller chipset is an ICH10R - but that is another question).

For those of us making do with these fakeraid cards on Windows-based platforms: How do you monitor your RAID array?

(If this is just handled by Windows, could you please point me towards a KB or TechNet article? I'm primarily an Unix admin, so if there is a simply a Windows equivalent of mdadm or raidtools I am ignorant of it).

---

Edit:

Things don't quite work:

• Using Virtual Disk Service via MMC doesn't work as a complete solution for me. It only sees the RAID-1 volume as it is presented to the operating system as a single disk - it has no knowledge of the underlying RAID. What I need is actual information about the health of the RAID itself (e.g., whether or not the array is degraded) not the volume it presents.
• I have installed the SNMP service and SNMPInformant, but I cannot find any reference in the MIBs for monitoring the ICH10R controller. This is an excellent suggestion but I need a little more detail to proceed. And again, how is the operating system supposed to know anything about the RAID? It just sees the abstracted RAID volume as a disk drive and the controller.
• I realize that the best I might be able to get is to use RDP to login and run Adapter Storage Manager locally but that is precisely the kind of thing I want to avoid. I can't be the only one who is stuck with using these fakeraid cards; I would hope there is a better solution for monitoring.
• The installer for Intel Storage Matrix Manager helpfully exits with the error: "This computer does not meet the minimum setup requirements for using the software". I presume this to mean that it is not designed to work with this controller (although it is listed as a supported device in the SATA ACHI section of the readme.txt).
• The Adaptec Storage Manager sorta works but doesn't really get me where I want to be. It doesn't "see" the individual drives that make up the array. It can however be 1) monitored remotely, 2) send email notification and 3) send SNMP traps. This is all nice and dandy, but if doesn't have the ability to interface with the controller it's not any more sophisticated in its monitoring than Windows' Virtual Disk Services management (see the screenshot).

Solution 1

Just some ideas:

You could try this software if the drives support SMART: http://www.panterasoft.com/ (see HDD monitor)

You could also get the Intel Storage Matrix Manager = http://downloadcenter.intel.com/Detail_Desc.aspx?agr=Y&DwnldID=17882&keyword=ich10r+console&lang=eng

I haven't used it for email notifications but I believe it and the RST console have that email option.

Or like Mike said, since the Storage Matrix manager fires off events to the eventvwr, you could use the feature in 08 R2 to grab those events and forward them via email or since you use OpenNMS you can grab the events http://www.opennms.org/wiki/Windows_Event_Log_Traps.

Solution 2

actually intel storage matrix software would be your best bet....

but you would need to first go into the BIOS and change your controller from RAID>ADAPTEC to RAID>INTEL (this will brick your current install of server08, yes that means a fresh install)

some open source software has a hard time installing on the intel fake raid so supermicro (and a few others) have made agreements with adaptec use their raid bios.

little late, I know. but i try to post answers when i pull threads from google (I'm actually looking for something completely different)

Solution 3

I think your first step is to locate what information is available within Windows about this array, or about the physical drives. I'd imagine that the logical drive appears as a single physical drive to the OS?

Use one of these tools to see if any information about this array or the physical drives has been exposed to the OS. - lodctr.exe - Typeperf.exe
- WMI Counters - HKEY_LOCAL_MACHINE\HARDWARE\DEVICEMAP\Scsi

Microsoft has WMI Explorer tools, but this is kind of a needle in the haystack using a gui. If you can get a VBscript to dump the classes (Rob Vanderwoude has one here), you can use WMIC.exe using the command line format of wmic.exe path [wmi path] get /format:list

If you can locate information about the drives or the array, you should be able to figure out a way to update Opennms.

Another thought: Have you ever simply removed one of the drives from the running system? Did it update the system event log in any way? Did the registry change, perhaps around key HKEY_LOCAL_MACHINE\HARDWARE\DEVICEMAP\Scsi?

Solution 4

I have three answers (or maybe suggestions would be a better way to put it).

First is that Windows MMC will allow you to remote into the machine and connect to the Virtual Disk Service on that machine. This would let you look at any windows managed volumes. Though this probably won't help you as only the Windows accessable Virtual Disk will show up, not the Physical Disk as managed by the RAID controller. But Remotely managing the machine will let you into the event manager as well and Adaptec's drivers do "tend" to write to the event logs, so you would have a place to go poke around if needed.

Secondly would be to look into the SNMP feature in 2008. You might discover that there is a MIB for monitoring that controller in there somewhere. (It's worth a look). You can turn on the SNMP Service under "Server Manager"->"Features->Add Feature->SNMP Services. It will install it as a system service which can them be configed under "Services" in Administrative Tools.

Third, remote desktop is your friend. I realize this isn't a good way to get a notification on disk failure, but it will let you login to locally look at the drive configs with something like Adaptec Storage Manager.

Just thought of a 4th option. IN 2008 R2 there is a feature in Event Logs that allows you to run apps/send emails based on thrown events in windows. You could use this to notify you of a failure.

Solution 5

Unfortunately, there is no standard way to watch RAID on Windows. So the best you can do is figure out how you (human) would do it, and then see if it can be automated.

Some systems write to the Event Log, and some have an embedded web page where you can monitor the current RAID status. I don't know about your specific hardware, but if either of the above is the case, then most server monitoring products would work. The free version of PA Server Monitor for example can monitor both of those and send alerts.

View more solutions

Comments

[Will Dixon]

The Adaptec-based controllers have a storage manager available; not sure about the Intel ones though.

[Nils]

Do you have a BIOS-option that allows you to pass the physical drives as well to the os?

[Admin]

The installer for Intel Storage Matrix Manager exits with the error: "This computer does not meet the minimum setup requirements for using the software".

[kralyk]

Someone else found this driver: downloadcenter.intel.com/… and was able to install it and get it to monitor at that point. YMMV be careful as it could hose the RAID array.

[kralyk]

Also, what is the software from Super Micro called "Super Doctor III"? I'm not familiar with their stuff overall, but it does say it comes with it and can monitor hardware. It might be worthwhile to call Super and ask them if there is a way to monitor it with their software.

[Admin]

I haven't gotten around to just simply pulling a drive out and seeing what happens -- in theory it shouldn't break but I'd rather not rebuild this server a second time... on the other hand it is better it breaks here than on-site.

[RobW]

Right. The point to pulling the drive is to see if windows notices a change in hardware. Thinking about this, the change could be as little as an increase in average seek or cache queue. If you can see something measurable, you can engineer a way to watch for it.

[Nils]

Well - there is a standard way using Linux - all "crappy" raid-controllers are supported through the same driver...

[Sergey Vlasov]

Using Intel Matrix Storage Manager would require switching from Adaptec to Intel firmware. This is possible on X8DTL-i and other similar Supermicro boards (look for the "ICH RAID Code Base" setting in BIOS setup), but changing this setting will destroy your existing array.

[kralyk]

Did you ever look into the Super Doctor III software? I can't tell if it does RAID info, but it appears to be their version of a WhatsUp or IBM Director, etc.

[Admin]

The server went out to Production a month or so after I posted this but I got the chance to work on earlier this month. Super Doctor III is the software I was looking for. It looks like it's SuperMicro's equivalent of IBM's ServerGuide or Dell's OpenManage. Unfortunately, while it supports hardware monitoring and SNMP alerts it does not support running on Server Core.