Did you try turning off this VM and rebuilding it from scratch - new VM, not a clone or from template?
Minisforum MS-01 PCIe Card and RAM Compatibility Thread
- Thread starter Patrick
- Start date
Notice: Page may contain affiliate links for which we may earn a small commission through services like Amazon Affiliates or Skimlinks.
Not yet, but I'm thinking that might be worth a try, just rebuilding exchange is annoying. I'm trying a few other things first, i've been seeing some repeating messages in my host logs that i've been working on isolating and cleaning up.
I noticed some osd and ceph-mon and ceph-mgg crashes.
I'm getting a message in my ceph logs on all my nodes that i find puzzling, and can't find much mention of with google
Code:
May 11 15:01:12 pve3 ceph-osd[2413]: 2025-05-11T15:01:12.993-0400 74fd475866c0 -1 bdev(0x5d57448c0c00 /var/lib/ceph/osd/ceph-2/block) aio_submit retries 2
May 11 15:01:54 pve3 ceph-osd[2413]: 2025-05-11T15:01:54.144-0400 74fd4257c6c0 -1 bdev(0x5d57448c0c00 /var/lib/ceph/osd/ceph-2/block) aio_submit retries 1
May 11 15:01:59 pve3 ceph-osd[2413]: 2025-05-11T15:01:59.743-0400 74fd465846c0 -1 bdev(0x5d57448c0c00 /var/lib/ceph/osd/ceph-2/block) aio_submit retries 3
May 11 15:01:59 pve3 ceph-osd[2413]: 2025-05-11T15:01:59.744-0400 74fd46d856c0 -1 bdev(0x5d57448c0c00 /var/lib/ceph/osd/ceph-2/block) aio_submit retries 4
May 11 15:01:59 pve3 ceph-osd[2413]: 2025-05-11T15:01:59.744-0400 74fd40d796c0 -1 bdev(0x5d57448c0c00 /var/lib/ceph/osd/ceph-2/block) aio_submit retries 4Attachments
Is it possible to do the BIOS 1.27 update without using windows? I don't see the a versioned .bin file like there is in the 1.26 download.
Nevermind, I found you can use the 1.26 UEFI download and put the 1.27 AHWSA.bin file on the USB as well.
Instead of running the AfuEfiFlash.nsh command, run the AfuEfix64.efi command instead, but specify the 1.27 AHWSA.bin file instead and the appropriate command switches as follows:
AfuEfix64.efi AHWSA.bin /p /b /n /x /r /l /reboot
Hello,
within this discussion some of you use a "GLOTRENDS PA09-HS" and "chenyang Oculink SFF-8612 zu PCI-E 4.0 NVME M.2 M-Key Host Adapter" to connect a egpu dock to your MS01 pcie port.
I bought the same components to connect the MS01 via oculink to a AOOSTAR AG02 EGPU Dock - but it does not work.
The dock & graphic card does not start up after switching on the dock & mini pc.
The red led on the pcie card lights up. No issues with system boot into proxmox.
Does anybody have some hints?
Do you made any special settings in the BIOS?
Do you made the SMBUS hack on the pcie card (I think not necessary, because boot process is fine)?
Is the adapter shown somehow in "lscpi" output - in my case no new device is listed.
Thanks for some feedback.
Dirk
within this discussion some of you use a "GLOTRENDS PA09-HS" and "chenyang Oculink SFF-8612 zu PCI-E 4.0 NVME M.2 M-Key Host Adapter" to connect a egpu dock to your MS01 pcie port.
I bought the same components to connect the MS01 via oculink to a AOOSTAR AG02 EGPU Dock - but it does not work.
The dock & graphic card does not start up after switching on the dock & mini pc.
The red led on the pcie card lights up. No issues with system boot into proxmox.
Does anybody have some hints?
Do you made any special settings in the BIOS?
Do you made the SMBUS hack on the pcie card (I think not necessary, because boot process is fine)?
Is the adapter shown somehow in "lscpi" output - in my case no new device is listed.
Thanks for some feedback.
Dirk
UPDATE: I'm pretty certain it's faulty, returning it to Amazon and ordering a new one.
I just got a MS-01 today and having issues with memory.
Should a single memory module work in either slot? I know some systems require single-modules to be installed in a specific slot.
I'm asking because I cannot get my MS-01 to boot when only the B1 slot is used or when both are used. When I have the memory in slot A1 it boots up fine using either of the two modules. As soon as I put memory in the B1 slot the system doesn't boot and never get anything out the HDMI.
I'm trying to figure out if it's an issue with the memory slot or potentially a memory incompatibility problem. The memory is a Mushkin MRA5S560LKKD16GX2 32GB kit (2 x 16gb).
I just got a MS-01 today and having issues with memory.
Should a single memory module work in either slot? I know some systems require single-modules to be installed in a specific slot.
I'm asking because I cannot get my MS-01 to boot when only the B1 slot is used or when both are used. When I have the memory in slot A1 it boots up fine using either of the two modules. As soon as I put memory in the B1 slot the system doesn't boot and never get anything out the HDMI.
I'm trying to figure out if it's an issue with the memory slot or potentially a memory incompatibility problem. The memory is a Mushkin MRA5S560LKKD16GX2 32GB kit (2 x 16gb).
Last edited:
Hi all, I tested with 50TB both Qnap cards, QXP-1600eS-A1164 (very noisy) and QXP-800eS-A1164 (very hot). Passthrough no problem, performance regular, and lot of problems with data corruption etc etc...
Same test with a SAS PCIe 3.0 SAS 12GB 24 Ports internes. Modèle OEM 9305-24i, cero problems, superfast, superstable no errors...
QXP-800eS-A1164 | Hardware Specs
I have the latest BIOS 1.27
Running proxmox 8.3.0 and doing the passthrough with All functions, now testing without all functions...
Any idea about what could be? I only have a few days before return the TL-D1600S JBOD storage
Thanks!
When you connected an NVMe SSD to the M.2 E-key (WiFi slot) did you get PCIe 3.0 x2 speed or PCIe 3.0 x1 ?
ASMedia chipsets that QNAP has chosen to use are OK for light disk usage, but if you're attempting to use them in a ZFS/TrueNAS setting, the chipsets buckle under the load quite quickly. TrueNAS forums will strongly recommend against their usage.
I was unable to make that 8-bay JBOD enclosure work even with an LSI 9207-8e, since I discovered that the enclosure uses the same ASMedia 1164 as the included QNAP SATA card and was still running into ZFS pool errors. I went through a very similar testing phase with the 8-bay QNAP and detailed it here:
Proxmox 8.3 + TL-D800s + LSI 9207-8e
Hi, so I'm pretty new here... My current setup is: MS-01 TL-D800s LSI 9207-8e Virtualizing TrueNAS Scale in VM under Proxmox with PCI passthrough. for now i'm populating the JBOD with 1 drive (in transition from DSM to TrueNAS) - my proxmox recognizes the drive, but also it lists ALL other...
forum.proxmox.com
In short, I settled on an LSI 9300-8e card, an iStarUSA/Raidage enclosure, and half meter (this is important) SFF-8644 cable to connect them.
@JaxJiang or anyone else who might know more about the thunderbolt ports.
As I understand the 13900h has on die thunderbolt controllers. What I don't understand is how much bandwidth do they have? Do the controllers use pcie lanes? if so do they share the pcie lanes?
I am using thunderbolt-networking to connect multiple ms01's. I am running network benchmarks and have found that I can get full bidirectional bandwidth (roughly 26gbps tx, and 26gbps rx concurrently) using only a single port. either port 1 or port 2.
If i try to use both ports at the same time, both with bidirectional tests, I get dropped packets, crc errors, and reduced bandwidth, if i set up rate limiting i get a fairly stable connection around 15gbps tx and 15gbps rx out of port 1 and port 2 at the same time. So the aggregate TX is about 30gbps, and the aggregate rx is also about 30gbps.
What i'm wondering is if both ports share a PCIE4 x4 link, which is roughly 32gbps full duplex which would explain my findings.
As I understand the 13900h has on die thunderbolt controllers. What I don't understand is how much bandwidth do they have? Do the controllers use pcie lanes? if so do they share the pcie lanes?
I am using thunderbolt-networking to connect multiple ms01's. I am running network benchmarks and have found that I can get full bidirectional bandwidth (roughly 26gbps tx, and 26gbps rx concurrently) using only a single port. either port 1 or port 2.
If i try to use both ports at the same time, both with bidirectional tests, I get dropped packets, crc errors, and reduced bandwidth, if i set up rate limiting i get a fairly stable connection around 15gbps tx and 15gbps rx out of port 1 and port 2 at the same time. So the aggregate TX is about 30gbps, and the aggregate rx is also about 30gbps.
What i'm wondering is if both ports share a PCIE4 x4 link, which is roughly 32gbps full duplex which would explain my findings.
I get 1x on the SSD that's connected to the E-Key slot.
Code:
LnkCap: Port #0, Speed 16GT/s, Width x4, ASPM L1, Exit Latency L1 <8us
ClockPM+ Surprise- LLActRep- BwNot- ASPMOptComp+
LnkCtl: ASPM Disabled; RCB 64 bytes, Disabled- CommClk+
ExtSynch+ ClockPM+ AutWidDis- BWInt- AutBWInt-
LnkSta: Speed 8GT/s (downgraded), Width x1 (downgraded)
TrErr- Train- SlotClk+ DLActive- BWMgmt- ABWMgmt-Do you intend to use this SSD as just a drive or as part of a RAID? If I understand correctly, a RAID pool speed cannot exceed the speed of its slowest drive. So if you do use it in a RAID, the entire pool will be limited to PCIe 3.0 x1 speed. Feel free to correct me on this one.
I use it as the boot drive and maybe some temporary local storage if needed. I don’t use a raid setup for my boot drive. I’m not entirely sure about the raid performance, but that’s probably likely unless maybe it can be tuned to use that drive less.
I don’t have large data requirements and don’t run raid anywhere except for a das with snapraid for some backups. Using the key E as the boot drive lets me put at least 16TB of storage in it using the other three slots.
Got a TXB122 for 2 extra m.2's
But this card hangs during a large file copy. At roughly 500GB the copy job just freezes. When cancelled, I find that some files are corrupted and the disk needs repairing. Running the same copy job to the same m.2 in a MS-01's slot works without issues.
Anyone else have problems with this card? Is it a bug? Defective? Or does the ASMedia chip gets to hot?
But this card hangs during a large file copy. At roughly 500GB the copy job just freezes. When cancelled, I find that some files are corrupted and the disk needs repairing. Running the same copy job to the same m.2 in a MS-01's slot works without issues.
Anyone else have problems with this card? Is it a bug? Defective? Or does the ASMedia chip gets to hot?
QNAP QDA-U2MP uses the same ASMedia 2812 chip and it didn't hang on me yet, so the issue is probably not the chip itself.Got a TXB122 for 2 extra m.2's
But this card hangs during a large file copy. At roughly 500GB the copy job just freezes. When cancelled, I find that some files are corrupted and the disk needs repairing. Running the same copy job to the same m.2 in a MS-01's slot works without issues.
Anyone else have problems with this card? Is it a bug? Defective? Or does the ASMedia chip gets to hot?
If you suspect that the thermal solution is not adequate, you could point a fan directly on the heatsink, to see if it solves the issue.
Yeah, I wanted to avoid the same issue. I have them all connected via cables, so I didn't need wifi or bluetooth. For home assistant type things, I use bluetooth proxies. Hope the e-key converters work out for you.
Update on my last post. PVE2 most was the node having unexplained ceph mon/mgr/osd crashes, and unexplained hard lockups that were happening pretty much daily.
I had a spare ms-01 that i bought for another project that I never set up, with new ram, and it's been over 24 hours without a crash, which so far is a record. I'm going on vacation for a week, so hopefully this was the issue. I'm assuming this is 13th gen degredation, but i'll test it out by swapping the ram once i get back from vacation. If so, will minisforum still warranty the device? it's about a year and half old.
I had a spare ms-01 that i bought for another project that I never set up, with new ram, and it's been over 24 hours without a crash, which so far is a record. I'm going on vacation for a week, so hopefully this was the issue. I'm assuming this is 13th gen degredation, but i'll test it out by swapping the ram once i get back from vacation. If so, will minisforum still warranty the device? it's about a year and half old.
A fan directly above the heat sink did not help.
However I suspect this only happens when using pci device passthrough. PCI mapping might have issues with the ASMedia chip, as this is more or less a workaround to present 2 nvme disks (devices) via a non-bifurcation pcie slot to the system.
I have tested it with a Windows VM and device mapping (which fails). Now I have 1 nvme formatted as NTFS, mounted on the host and have copied over a 100GB without issues. Will try with a linux VM and device mapping, to see if this will fail too...
Last edited: