Intelligent Enclosure - SAF-te Info

[CA1295230]

Apparently several of my servers use something called SAF-te. I've done some researching and find that this is some sort of SCSI interface to the actual rack/enclosure. It allows monitoring of fans, etc. How can I find out more about the SDR GEM318 controller or processor that is in use? I'm assuming this was once supported by HP. Where can I get the driver? Instant Toptools says it's located as Controller 2, Channel 0, Target 11. Thanks.

[kris rombauts]

Hi Dane,

can you specify if you see any 'new devices detected messages' at bootup of your server (assuming you run Windows) and if you are talking only about internal disk backplanes or also external SCSI disk enclosures like RS/12 which are connected to your server i.e. ?

Their isn't any special driver needed that you need to load except for some you might need a dummy .inf file to properly recognize it in Windows device manager (to get rid of the yellow question mark) but other then that their is no driver really.

When temperature events, disk power issues, disk inserts, disk removals etc take place in the disk backplane, it will report the problem and it will be logged by the management agents (TopTools agents) into your System log file at the operating system level i.e.

I am just puzzled with the ID=11, normally it is at ID=5, so describing your configuration a bit more might help, do you have a Netraid 1M or 2M as the raid controller or the embedded ctrl i.e. ?

Under the Windows device manager, system devices you should see a "HP SAFTE SCSI processor device" to represent that device.


HTH

Kris

[CA1295230]

Hi, Kris.

No, I do not see "new devices detected" when I boot up. This server is hooked to a RAID-5 array via MS cluster, so I don't actually see the RAID array unless the cluster has failed over to this node (secondary node). I'm not sure about the internal disk backplane: I have two hard drives in the server (NetServer LP 2000r). Yes, Device Manager sees it as a Question Mark. I have read HP docs on how to get it to recognize it. However, clicking on the Question Mark displays it as "SDR GEM318 SCSI Processor Device." I may not bother with the .inf file if it's not critical. But obviously, Device Manager does not see it as "HP SAF-te SCSI Processor Device" now. I am running the NetRAID-2M controller. I'm struggling to try to rebuild this server, to put it back as the secondary node of the MS cluster. I finally joined it to the cluster yesterday, but in doing do, must have overwritten or somehow lost the Instant TopTools software that showed the configuration. I'll get busy putting that back on the box. I could be that I'm not sure which software to use to take advantage of this SCSI interface that can report events that you describe. Any clues? Is it more than Instant TopTools? Thanks.

[kris rombauts]

Dane,

two things to do or verify:

- load the .inf file to get rid of the question mark in dev. mgr
The .inf comes with the Netraid 2M driver package.

- make sure the Netraid monitor service is running, this will also contribute to reporting any errors detected by the SAF-TE processor and by the Netraid controller like disk drive errors, logical drive issues, etc etc.
The Netraid monitor service is available here:


http://h20000.www2.hp.com/bizsupport/TechSupport/SoftwareIndex.jsp?lang=en&cc=us&prodNameId=62488&prodTypeId=329290&prodSeriesId=51930&swLang=13&taskId=135&swEnvOID=226


HTH

Kris

[CA1295230]

Thanks, Kris.
I've fixed the "?" with the inf file. I'll pull down the "monitor" software. I notice there are several choices in this area. Which is recommended?

[kris rombauts]

Hi Dane,

download the last monitor version that ever released (V 5.07) and let it install with defaults so that it also performs a weekly consistency check on your raid subsystem.
This is very helpfull to make sure the raid/parity information is consistent.

Also check if the Netraid 2M runs the last F/W H.02.02 since their is an issue with H.02.01 with inability to rebuild when their are bad blocks on a disk and you don't want to run into that issue whenever a single disk gets some bad blocks.


HTH

Kris