udma_crc_error_count of 820.

[brianm]

Having carried out many data transfer tests on my new system I thought it was time for some baseline measurement tests.

I ran Smartcl and everything, apart from the following two items, were as you would expect. These two items were both way out of parameters on all 4 drives. I ran Scrub with no change.

The drives are 4 new WD Red 2TB in a 2x2 mirror. The disk controller is an H200 patched to 20.00.07.00. Freenas is 11.0.U4.
The 305 watt power supply is driving the 4 drives, an E3-2270 CPU and the hard disk controller.

These are the problems reported.
Spool up time on all drives is around 4500.
udma_crc_error_counts are all around the 800 mark.

I replaced the disk controller cable with a second (also used) part but the crc counts did not change. Since all 4 drives are showing the same indications and I tried a replacement controller connection I am thinking that the H200 is bad or perhaps the power supply is failing. Reflashing and replacing both of these parts is going to be time consuming since I do not have spares on hand.

Before I start the replacement route what do the experts think?

Thanks, Brian

 

[Stux]

Udma CRC count never resets, so if it’s not going up you’ve resolved it.

Fwiw, it almost always indicates a faulty cable

 

[brianm]

Thanks Stux, I will monitor it.

 

[Ericloewe]

Definitely replace the cable.

 

[brianm]

Thanks Ericloewe,

I have the replacement cable (used but better looking) in place now so I will start a log of the counts. Stux provided a key because I did not realize that the count was a "life to date" value.

Direct connection of the drives to the motherboard SATA connectors had previously worked fine so it was clear that the problem was in the controller to disk chain.