Hello,
After upgrading Proxmox Backup from 4.0 to 4.1.x and switching the kernel from 6.14.11-6-pve to 6.17.13-2-pve, the drives connected to the DIGITUS DS-30104-1 controller (Marvell 88SE9230 chipset) are no longer detected.
Hardware details:
I want to highlight that this issue occurs only after switching to kernel 6.17.13-2-pve. I hope the developers can look into whether future kernel updates can restore support for Marvell 88SE9230 controllers, or if support for this hardware is being deprecated.
This appears to be related to the newer kernel version. Full log files (lspci, dmesg, lsscsi, journalctl) are available as attachments for further debugging.
Thank you in advance for looking into this issue and any suggestions.
After upgrading Proxmox Backup from 4.0 to 4.1.x and switching the kernel from 6.14.11-6-pve to 6.17.13-2-pve, the drives connected to the DIGITUS DS-30104-1 controller (Marvell 88SE9230 chipset) are no longer detected.
Hardware details:
- Server: Dell PowerEdge T110 II
- Motherboard: 015TH9
- Controller: DIGITUS DS-30104-1, Marvell 88SE9230
- Kernel driver: ahci
- Drives were fully detected on kernel 6.14.11-6-pve.
- After booting kernel 6.17.13-2-pve, lspci -vvv -s 02:00.0 shows the PCIe link downgraded from x2 to x1.
- Other SATA controllers on the system continue to work normally.
- Logs from lsscsi, dmesg, and journalctl confirm that the drives are not recognized.
- A known issue with Marvell 88SE9230 (https://forum.proxmox.com/threads/proxmox-cant-recognize-hdd-on-sata.60259/post-668900) suggests adding iommu=pt to GRUB to resolve detection problems.
- In this case, applying iommu=pt did not solve the problem.
I want to highlight that this issue occurs only after switching to kernel 6.17.13-2-pve. I hope the developers can look into whether future kernel updates can restore support for Marvell 88SE9230 controllers, or if support for this hardware is being deprecated.
This appears to be related to the newer kernel version. Full log files (lspci, dmesg, lsscsi, journalctl) are available as attachments for further debugging.
Thank you in advance for looking into this issue and any suggestions.