Log in

No account? Create an account

Josh-D. S. Davis

Xaminmo / Omnimax / Max Omni / Mad Scientist / Midnight Shadow / Radiation Master

Previous Entry Share Next Entry
SATA fakeraid
Josh 201604 KWP
So, I found that the RAID controller is technically BIOS/Software RAID. I can deal with that.

The problem is that the raid drivers for linux for this are a devicemapper plugin (dmraid) which is alpha. It shows degraded mode because I didn't format before use. It doesn't support updating state in the metadata yet.

Oh, and it doesn't obscure the backing devices in any way, so I'm concerned about Linux LVM, at some point, getting confused, or maybe not even being able to find the RAID set and just using the raw devices.

So, I think I need to disable the BIOS metadata tagging of he devices and just use the linux MD drivers.

  • 1
I've tried countless times to deal with raid controllers that work that way, and every time I've come to the conclusion that, if it isn't "real" raid, just use MD or get a real RAID and pay close attention to modules when you change kernels.

IMHO, HW raid should pass I/O errors up through the LUNs. I don't know how I/O errors are differentiated at the kernel level, but there should be a way to report them without causing a retry or failure.

Effectively, something that says it was retried and succeeded, or something that indicates a degraded mode.

The OS should pass these through to syslog, or equivalent.

Having to deal with device drivers for something like this is asinine.

  • 1