Poweredge R610 fan management on debian

Servers

Servers
Information and ideas on Dell PowerEdge rack, tower and blade server solutions.

Poweredge R610 fan management on debian

This question is answered

Hi,

I have this R610 server with Dell OpenManage packages installed, srvadmin-*!

Now.. the server is incredibly LOUD because of the cooling system and no matter tha preset I choose on BIOS (Maximum performance, OS or Active demand) the coolers are always on 8000+ rpm.

I've been googling all day to see if I can get to manage this but no luck, when I try to set the value using OpenManage cli tool I get:

root@test:~# omconfig chassis fans index=0 warnthresh=default
Error! Default fan threshold setting unavailable for index: 0.

And this is how the fans are looking like now:

root@test:~# omreport chassis fans
Fan Probes Information

Fan Redundancy
Redundancy Status : Full

Probe List
Index : 0
Status : Ok
Probe Name : System Board FAN MOD 1A RPM
Reading : 9840 RPM
Minimum Warning Threshold : [N/A]
Maximum Warning Threshold : [N/A]
Minimum Failure Threshold : 1920 RPM
Maximum Failure Threshold : [N/A]

Index : 1
Status : Ok
Probe Name : System Board FAN MOD 2A RPM
Reading : 9840 RPM
Minimum Warning Threshold : [N/A]
Maximum Warning Threshold : [N/A]
Minimum Failure Threshold : 1920 RPM
Maximum Failure Threshold : [N/A]

Index : 2
Status : Ok
Probe Name : System Board FAN MOD 3A RPM
Reading : 9720 RPM
Minimum Warning Threshold : [N/A]
Maximum Warning Threshold : [N/A]
Minimum Failure Threshold : 1920 RPM
Maximum Failure Threshold : [N/A]

.. And so on, all the same.

Please, I'm completely lost, if I can't manage this with dell's official OpenManage tool, then what?

Thanks

Verified Answer
  • I'm no PE expert but IIRC, OpenManage allows you to simply see what the fans are doing and to manage the notification settings for when fans behave outside their designated speeds. That is, you can set at what speed you will be notified of an issue or at what speed the server will be automatically shut down - nothing more - IIRC.

    I'm also not sure at what speed a 'normal PE R610 spins it's fans but it all depends on what is installed inside the chassis, the CPU load and the room temp. If these details are spelt out, maybe someone can indicate if your fan speeds are within "normal" for your R610 or if you might have a h/w fault that should be addressed.

    In reality, Dell has designed the PE cooling strategy based on their view of how their machines should be used (in cooled server room) and based on the warranty they provide (years). As such, Dell provides no tools or methods to manage the logic of how the installed equipment inter-operates with the fan speed which inter-operates with the temperature sensors. This logic is within the BIOS/BMC and it's a no go area for owners, you just can't play with these preset curves/settings. BIOS and BMC firmware updates may also help to quiet down your machine if Dell has improved/changed the logic....

    Having said all this, some few have found hacking their BMC/BIOS and installing slower fans a workaround in their old unsupported machines, see this project page and this project page for some insight on what may be involved should you want to go down this unsupported path....

All Replies
  • I'm no PE expert but IIRC, OpenManage allows you to simply see what the fans are doing and to manage the notification settings for when fans behave outside their designated speeds. That is, you can set at what speed you will be notified of an issue or at what speed the server will be automatically shut down - nothing more - IIRC.

    I'm also not sure at what speed a 'normal PE R610 spins it's fans but it all depends on what is installed inside the chassis, the CPU load and the room temp. If these details are spelt out, maybe someone can indicate if your fan speeds are within "normal" for your R610 or if you might have a h/w fault that should be addressed.

    In reality, Dell has designed the PE cooling strategy based on their view of how their machines should be used (in cooled server room) and based on the warranty they provide (years). As such, Dell provides no tools or methods to manage the logic of how the installed equipment inter-operates with the fan speed which inter-operates with the temperature sensors. This logic is within the BIOS/BMC and it's a no go area for owners, you just can't play with these preset curves/settings. BIOS and BMC firmware updates may also help to quiet down your machine if Dell has improved/changed the logic....

    Having said all this, some few have found hacking their BMC/BIOS and installing slower fans a workaround in their old unsupported machines, see this project page and this project page for some insight on what may be involved should you want to go down this unsupported path....

  • Ok,

    Thanks for your complete answer, it is really clarifying. It's my first time with dell servers so...

    Ok, if then I can't manage the fan speed, then how can I update the correct firmware that manages that? On Linux? How can I update specificly the BMC firmware? I've downloaded the following files but I'm not sure how to do it.

    SUU_730_x64_Q22013_A01.iso

  • To be more concrete:

    http://www.dell.com/support/drivers/us/en/04/driverdetails?driverid=J5GTM

    Here I have a firmware version for BMC. How do I know it's for R610, since when I get to R610 drivers & downloads I can't find any BMC downloads? I figured that may be R610 never has a BMC update (which is odd) and also can't find BMC firmwares for other servers.

  • I was able to update the firmwares using SUU after mouting the ISO on linux.

    I've had to install libstdc++6, rpm, openjdk7 and link /bin/sh to /bin/bash to so SUU command work.

    The server is incredibly noisy though.. I'll open a new thread asking questions about the fan connectors.

    Thanks

  • @xgonzalo, i'm no PE Server expert as mentioned earlier.. BMC stands for Baseboard management Controller which i think is an older term for the 'object' that managed motherboard features along with the BIOS... On the R610, i think the 'object' that handles sensors, fans and other management functions for the server is the iDRAC, Integrated Dell Remote Access Controller (listed under ESM section on the R610 driver downloads page) !! It is this firmware that you may need to updated...

    With the BMC firmware link you referenced in your previous post, and any dell download page for that matter, there is a Compatibility section contained within but in this case it only lists 'Enterprise server 2900' & 'PowerVault DP600/NF600' as compatible with this BMC firmware, your R610 is not listed as compatible with this BMC firmware!! Be careful to always check the compatibility section to ensure the download applied to your hardware...

    Now on your other thread on this issue, I've provided a reference to the 'dell drivers and downloads page' for your R610 which you should browse for relevant downloads for your system. And since you're new to PE servers, read the 'Technical Guide Book' and 'Hardware Owners Manual' for your system, again look at your other thread for links...

    Sorry i didn't catch this earlier but at least it seems the BMC firmware update did no damage Embarrassed

    And unfortunately Dell 1U servers will always be loud...