ProLiant DL20 Gen10 very noisy

[sudhirsingh]

Hi,

Please provide hardware information (take screenshot from all the tab under ILO->system information)

Suggested RBSU setting:

Work load profile-> HPC (High performance computing/Virtualization Max performance)

Collaborative power control->Disable,

Cooling mode -> Optimal cooling [ if same issue try changing to "Increased cooling Mode"]

Share status,

Thanks,

[Pierrot Robert]

Hi,

Collaborative power control->Disable : it is already disabled

Work load profile-> Virtualization Max performance : no change in noise

Here are the requested screenshots:

[sudhirsingh]

Hi,

Can you please share ahs log from the server.

Regards,

Sudhir

[mysy]

you need make sure your have connect some sata hdd to mian board sata port, if only use HBA/Raid card , some reason from old DL320e and DL20 , the system not read any temperature from hdd, it will keep the fans to run in higher speed.

For my case, I connect a 8GB ssd to the main board sata port and power it with ODD power with convert interface.

[sudhirsingh]

Hi,

Do you have any non HPE device connected (e.g ..any M.2/NVME drive...) if yes, try to remove them and check.

Can you share snap from ILO5->Power and Thermal (Need to check temperature recorded for all the devices)

Also makesure Agentless Management Services has been started and running, if not restart manually.

Regards,

Sudhir

[adkeric]

I'm running into the same issue here, I'm trying to cut some costs for a lab environment and therefore I didn't purchase the system with HPE drives in it, and I have a Barracuda SSD attached to the cabling for the onboard RAID controller.

I know from previously working with a bunch of HPE hardware that it would potentially crank the fans up crazy if there was a non-HPE component installed, but I'm looking to avoid having to dump a lot of personal money into this system.

To the point about connecting a drive directly to the SSD port, what part numbers would satisfy this cable? I looked at the parts manual and saw P07894-001 as the spare for the optical drive cable, which is an insane price of $130 USD on eBay, and even more from parts sellers. So then my mind goes to would I be better off buying some spare HPE drives and just using them, again, I would rather not spend a ton of money.

If I went and bought a couple of these drives, HPE 1TB 6G 7.2k rpm SATA (3.5in) Non-Hot Plug Entry 512e [843266-B21], I'm trying to determine if they would then have the sensors necessary to help control this fan issue (having read that it's a lack of temp reporting from drives earlier in this forum thread).

I'm running Windows Server 2019 on this system, and have the agentless management service installed.

I appreciate any insight into this, as I'm trying to share living space with this lab and cannot deal with the fan noise.

[Artstar]

I appreciate any insight into this, as I'm trying to share living space with this lab and cannot deal with the fan noise.

I certainly understand where you're coming from with this. While I understand and appreciate that all manufacturers want their customers to stick with qualified hardware in order to be able to provide reliable support of their product, I think there should be some leeway for those who wish to accept the risks concerning the support and use third party (aka unqualified) hardware as well.

In my case, Fan #3 is running at 6% because I've assigned the SmartArray SW controller rather than AHCI and it seems to be satisfied that since I don't have any drives plugged into any of the slots, there's nothing to cool. However, Fan #1 and #2 are running at around 71% because I am using a Samsung 970 Pro NVMe SSD in the M.2 slot at the rear of the motherboard.

Long and short of it, I don't believe there is any way around the cost saving exercises that you and I are attempting here. I even found a post from someone else on this forum who has a genuine HP NVMe SSD plugged in with the same problem anyway, though I would expect it was eventually sorted out despite the silence in that thread.

Educationally speaking, the only way out of it, as far as I can tell, is to use additional circuitry to modify the duty cycle of the PWM signal sent to the fans, thereby limiting the maximum speed at which they can operate, so long as you keep a close eye on all your system's temperatures. Until HPE can learn to show a little more flexibility to its customers who are prepared to take responsibility for their own actions, people will take matters into their own hands and deal with the consequences thereafter. Had it not been for the excellent form-factor of this server, I know I would be selling this thing off and looking to another small 1RU solution for my ESXi server instead.

[ChoodJr]

We currently use HP Proliant DL20 GEN10 servers. Previously used DL320 GEN8 and DL160GEN9 and problems with noisy fan was always addressed with a SPP or ILO update. Not the case with this model. And I have worked with HP trying to find a solution. For now I have decided to go with High-Performance Mode. Hope this helps...

Please upgrade the AMS service from existing old version to 2.30 per HP...

Link :  https://support.hpe.com/hpsc/swd/public/detail?swItemId=MTX_e9e17b6c3c3f46f58f738bcb3f

Testing: Observations about Power Regulator modes: All in a Custom Profile

Step 1: Change you Workload profile in bios to HPC then save.

Step 2: Change your Workload profile to Custom

Step 3: Change Power and Performance Options choose one of the following modes each will spin down the fans, but some take a longer time than others with a warm reboot.

      - Note: Collaborative Power Control should already be disabled if not disable.

Static High-Performance Mode = 10 minutes or more to spin down fans (Speed 40%) after a warm reboot. Max CPU speed (4.25 GHz - 4.25 GHz) in use Fans may, 1,2 fluctuate 18% Fan 3 stays at 6%.

Dynamic Mode = 10 minutes or more to spin down fans (Speed 40%) after a warm, hard or bios reboot . Max CPU speed (2.xx - 4.xx GHz) in use Fans may, 1,2 fluctuate 18% Fan 3 stays at 6%.

• Note 1: Cannot set Collaborative Power Control to disable the recommended setting while in this mode.
• Note 2: Whether a warm, hard, or bios reboot fans will always take 10 minutes or more.
• OS Mode = Same as Static High-Performance Mode…
• Static Low Power Mode = 2 minutes or more to spin down fans (Speed 40%) after a warm reboot (This mode behaves normal) Minimal CPU speed (1.xx – 1.59GHz GHz) This mode however seems to be too slow. I witnessed this while applying a cumulative Microsoft patch took over 15 minutes. Typical time 6 minutes.

Note: A hard shutdown such as a power off or a bios change with a custom Profile and one of these modes’ fans will spin down in 2 minutes or more. However, a warm boot will take 10 minutes or more. Except for Dynamic Mode. I speculate it’s because Collaborative Power Control is enabled. See Notes above for this mode.

[Nashon]

Installing the Ageless management Service solved my case.

https://support.hpe.com/hpsc/swd/public/detail?swItemId=MTX_e9e17b6c3c3f46f58f738bcb3f#tab2

Remember to install the prerequisite Channel Interface driver first

https://support.hpe.com/hpsc/swd/public/detail?swItemId=MTX_972cb06cbe6645758421ac9cb7#tab3