Problems using Nextcloud on DL380 Gen8 + P420 2GB FWBC and RAID 1+0

[octooptokoppler]

Hello,

as I am not totally aware of if this problem is more software related than hardware, I am also reaching out to this forum as I am running out of ideas and knowledge. I am running a HPE ProLiant Microserver Gen8 (Product ID 819185-421) I got from my brother. I upgraded the CPU from the Celeron 1610 to a Xeon E3-1280 V2 and installed a P420 controller card having a 2GB memory card and FWBC installed. My four 4TB HDD (Seagate IronWolf Pro) are configured as a RAID 1+0. As the P420 controller gets very hot, I installed a small 5V Sunon undustrial fan facing the heatsing and decreased idle temp from 85°C down to 53°C.

I am using Ubuntu Server 24.04.1 LTS running Nextcloud Snap 30.0.4snap2. In december 2024, I faced an overheating issue with the P420 (after that I installed the fan) and the RAID failed, luckily, I was able to re-enable the RAID and all data were restored. From that on, my Nextcloud crashed. After reinstalling it, everything is now up and running again.

Now to the problem: when setting up the Nextcloud desktop client, everything works until I want to synchronize large data or folders having more than 1 GB of size. System loads increses until the whole system freezes, the P420 disappears in the iLo Temperatures page (Sea of Sensors 3D), the Server Fan increases speed from 26% idle to 70% and Ubuntu shows up I/O failure on sda1 (which is the RAID), resulting in remounting it in read-only. From that on, I am unable to use Nextcloud until reboot of the whole server. I also tried adjusting the controller settings from standard to On-Demand Video and Write/Read Cache ratio from 10/90 to 50/50, but no improvement at all. I am also aware of this forum post (DL380e G8 + P420 - P420 crashing - Hewlett Packard Enterprise Community) but it does not help either.

Some specs:
- System ROM J06 04/04/2019
- iLO Firmware Version 2.82 Feb 06 2023
- Smart Array P420 Controller Product Version B; Firmware Version 8.32
- Intelligent Provisioning Firmware Version 1.62.14
- SPS Firmware Version 2.2.0.31.2
- System Programmable Logic Device Firmware Version 0x06

Thanks a lot in advance!

[shiva_jr]

Hi @octooptokoppler , 
     I feel that, you don't have the issue with hardware like server, disk drive or fan. Issue may be configurations of Nextcloud. However, try these steps -
1) Check the fans ventilation path is clear, no cables are stopping the air flow.
2) validate that all the replaced drives are supported with existing drive speed.
3) Make sure that the server will run properly if Nextcloud is disabled in OS.
4) Refer these posts  (Post1, Post2, Post3 ) for getting the solution if the issue is in Nextcloud. 
Regards,
Shiva_jr

[TVVJ]

Hello,

Third party hardware products are not optimized for iLO's sea of sensors and may not report temperature appropriately. iLO, based on its thermal sensors may require to turn up the fans to compensate against possible thermal issues.

Please install HPE option parts and see if the Smart Array controller functions in the appropriate temperature range.

Also, for the cloud solution, please ensure that the hardware is as recommended and the software is configured as recommended by the cloud solution vendor.

Regards,

[octooptokoppler]

@shiva_jr  Thank you for reaching out, I will check these posts and see if they can improve anything. Sice posting my question and not doing any huge data transfer, system is up and running smoothely.

@TVVJ Thank you, too, for reaching out. I am a bit confused to what you may refer with third party hardware? The server runs completely with genuine HP parts, except:
- RAM modules (not failing since when in service for my brother when he owned the server)
- SSD for OS (Kingston 120GB that was left from a former PC)
- four 4TB Seagate IronWolf Pro NAS HDD (bought in late 2024)
The RAID controller card P420 along with it's cache module and FWBC is a genuine HP part. The fan that I installed to cool down the main IC underneath the heatsink of this card is plugged into the USB port that's directly on the server's motherboard and only connects to +5VDC and GND.  When my actual issue comes up, the P420 completely disappears in the Sea of Sensors page, but all four HDD are shown correctly in the Storage page.
So, what HPE option parts do you recommend I may install?

[octooptokoppler]

So, I checked your recommendations, as follows:

---

@shiva_jr  schrieb:

1) Check the fans ventilation path is clear, no cables are stopping the air flow.

I checked that and all paths are as clear as possible. For better cooling, I 3D-printed a fan housing for the P420 card and a fan housing for the CPU heatsink do increase cooling and decrease temps. So far, working properly, even when performing stress tests.

2) validate that all the replaced drives are supported with existing drive speed.

---

==> All drives support SATA6 and have 7,200RPM

3) Make sure that the server will run properly if Nextcloud is disabled in OS.

==> It does. Only additional application is a Samba server for one of actually two attached USB3.0 external hard drives. Even this works properly.

4) Refer these posts  (Post1, Post2, Post3 ) for getting the solution if the issue is in Nextcloud. 

==> I checjed these posts, but none of them applied to my problem.

What I tried and found out is, that the P420 card seems to fail under heavy load. After all tests, I finally stopped nextcloud and copied my files folder (~90GB) to one of my external hard drives. After approx. 10GB, my initial failure occured again: the P420 disappeared in the iLO Sea of Sensors list.

Any suggestions?

---