PCI-e Sata expansion card

[sanfark]

2 hours ago, JorgeB said:

It's not dedicated to each port, if you have two disks connected, but are only sung one, it will have the full bandwidth.

Excellent. That being said, would be be unwise to over provision your PCIe x1 throughput by adding more disks than you have throughput? for example using a 4 port SATA controller on PCI 1.0.

 

If you leave your parity drive on MOBO dedicated SATA and your cache on MOBO dedicated SATA, one would believe it is fine to over provision or am I missing something? I am trying to think of a scenario where all disks are read/writing at once; then over provisioning would be unwise. Parity check perhaps?

[JorgeB]

7 minutes ago, sannitig said:

Excellent. That being said, would be be unwise to over provision your PCIe x1 throughput by adding more disks than you have throughput? for example using a 4 port SATA controller on PCI 1.0.

It's fine as long as you understand performance will suffer if they are used all together, you could have for example just one array disk there and use other ports for a pool, they wouldn't interfere if used separately.

[DerAndi]

Hallo zusammen,

 

ich glaube ich habe hier irgendwo schon im Thread jemanden gesehen der die IO Crest am laufen hat mit 5 Port.

 

O Crest Interne PCI-E x4 Controllerkarte

 

mich Interessiert natürlich wie lange diese schon im Einsatz ist und ob die einwandfrei mit Unraid funktioniert oder  ob die gar Probleme macht? Frage deshalb da ich bald eine Karte für die Erweiterung brauch.

 

gruß

[Gragorg]

13 hours ago, DerAndi said:

Hallo zusammen,

 

ich glaube ich habe hier irgendwo schon im Thread jemanden gesehen der die IO Crest am laufen hat mit 5 Port.

 

O Crest Interne PCI-E x4 Controllerkarte

 

mich Interessiert natürlich wie lange diese schon im Einsatz ist und ob die einwandfrei mit Unraid funktioniert oder  ob die gar Probleme macht? Frage deshalb da ich bald eine Karte für die Erweiterung brauch.

 

gruß

Diese Karte sollte gut funktionieren. Es handelt sich um einen unterstützten Chipsatz und eine anständige Marke.

[MrAndyBurns]

If it's of interest, I had to get 2 extra sata ports (Mobo/CPU upgrade with less ports then to find dead sata port) and was looking for a quick fix. I can't afford a nice HBA card so I went for an 'KALEA-INFORMATIQUE PCI Express PCIe x1 2 port controller card for SATA hard drives. Compact format, with ASM1061' as I believed the ASM1061 chip was okay ... sadly the card was not recognised by the system so I returned it and got a 'MZHOU PCIe SATA Card 2 Port - PCI Express SATA 3.0 GEN3 Expansion Card - 6 Gbp/s SATA Controller Expansion Cards with Low-Profile Bracket (Chip JMICRON+JMB582)'

 

So far so good with parity check running well for all drives and a drive passed through to a VM.

 

 

[asktoomuch]

On 12/1/2022 at 11:53 PM, tr3bjockey said:

UPDATE:  This has worked flawlessly and no more UDMA CRC error counts.  Going to toss this IO CREST Internal 5 Port Non-Raid SATA III in the trash which was causing my problems.

 

Are you sure your UDMA CRC issues were coming from the IO CREST JMB585 card? I see this specific card recommended in multiple threads on these forums, including this one.

 

I have been struggling for several years with UDMA CRC errors on my A88X plaftorm. My motherboard has 8 SATA ports, but whenever I go over 6 hard drives, it seems to become unstable and register a bunch of these errors on random drives. I checked the block diagram of the A88X chipset and the 8 SATA ports are supposed to be handle by the chipset itself, none is connected directly to the CPU, so that should rule out different behaviour between different SATA ports.

 

I have replaced the motherboard by a different model from a different brand (original was AsRock, now I have an Asus one), all the SATA data cables, the PSU, the SATA power cables... Nothing seemed to fix the problem. I confirmed that all my drives are fine too in a different machine.

 

So I gave up and for the past few years, I have used only 6 drives, despite owning a Pro license.

 

But last week, I decided to replace my two 8TB parity drives (shucked WD My Book) by two recertified 12TB HGST drives and again, I observed the UDMA CRC errors happening on most of my drives, including old 3TB drives. I was confused, because the new 12TB drives are basically identical to the old 8TB ones, specifying the same power requirements (400mA 5V, 550mA 12V).

 

So I have ordered an IO CREST JMB585 card from their official Aliexpress store (my local Amazon doesn't have it) and I'm hoping it will help. But your message makes me rather less hopeful than I originally was, if this card was causing you the same kind of issues that I face.

 

If that doesn't help, I might have to give up on Unraid completely, because at this point I have replaced pretty much everything except the RAM (Memtest verified) and CPU...