[mythtv-users] RAID 1 mirror with mythtv backend???

[James Abernathy]

On 1/26/19 9:41 PM, Stephen Worthington wrote:
> On Sat, 26 Jan 2019 09:26:24 -0500, you wrote:
>
>> On 1/25/19 6:37 PM, Will Dormann wrote:
>>> On 1/24/19 12:05 AM, Stephen Worthington wrote:
>>>> Those are CRC errors, and in a large quantity like that they are
>>>> normally indicative of a bad SATA cable (or one that is not firmly
>>>> plugged on).  Or possibly a USB cable problem in this case.  Dirty
>>>> connectors will also show up like this, and having the cat tread on
>>>> the cable can also do it.
>>>>
>>>> If the cable is OK, then the next most likely candidate for the cause
>>>> is a bad SATA port.
>>> I concur here.   From your SMART report:
>>>
>>>     5 Reallocated_Sector_Ct   0x0033   100   100   005    Pre-fail
>>> Always       -       0
>>> 196 Reallocated_Event_Count 0x0032   100   100   000    Old_age
>>> Always       -       0
>>> 197 Current_Pending_Sector  0x0022   100   100   000    Old_age
>>> Always       -       0
>>> 198 Offline_Uncorrectable   0x0008   100   100   000    Old_age
>>> Offline      -       0
>>>
>>> All of your raw values (the last column) are 0.   This implies that the
>>> drive itself is internally fine.  Which generally leaves the SATA
>>> controller and/or cable.
>>>
>>> Additionally from your report:
>>> SATA Version is:  SATA 3.0, 6.0 Gb/s (current: 1.5 Gb/s)
>>>
>>> I assume that your controller is only SATA 1.0?   If not, then that's a
>>> likely sign that your cable is bad enough that the drive negotiated down
>>> to a slower speed due to unreliable communications.
>>>
>>>
>>> -WD
>>
>> The controller is part of the Sandy Bridge chipset so it's 2 SATA 6GBs
>> and 4 SATA 3GBs.  So I moved the drive in question to another test PC
>> with same controller and new Cable, the locking kind. Then ran the
>> smartctl -t long and it came out complete good.
>>
>> I have also replaced my cables on the 2 locations in question in my
>> production system. No issues yet there.
>>
>> So I'm going to leave things untouched for a month and see what happens
>>
>> Jim A
> Drive SMART self tests do not move data off the drive to the PC, so
> they do not test cable and SATA port faults.  So the drive passing a
> long self test is another indicator that the drive itself is probably
> OK and the problem is elsewhere.

makes sense because the system under test was a different motherboard 
and cable. I ran the smartctl long test and Spintrite 6.  on that 
system.  Spinrite does read and write every sector using the PC so it 
checks all the parts of the system related to the drive.

The production system with the new cables has been quiet for a couple of 
days.

Jim A