Re: Seagate ST6000DM003-2CY186 "flapping". Disconnecting and reconnecting randomly.

On 2026-08-11 00:36, Jonathan wrote:
> On 2026-08-10 18:25, Daniel Lysfjord wrote:
>> On 2026-08-11 00:10, Jonathan wrote:
>>>>> 
>>>>> SMART Attributes Data Structure revision number: 10
>>>>> Vendor Specific SMART Attributes with Thresholds:
>>>>> ID# ATTRIBUTE_NAME          FLAG     VALUE WORST THRESH TYPE      
>>>>> UPDATED  WHEN_FAILED RAW_VALUE
>>>>>    1 Raw_Read_Error_Rate     0x000f   073   064   006    Pre-fail  
>>>>> Always        -       20195647
>>>>>    3 Spin_Up_Time            0x0003   098   092   000    Pre-fail  
>>>>> Always        -       0
>>>>>    4 Start_Stop_Count        0x0032   100   100   020    Old_age   
>>>>> Always        -       85
>>>>>    5 Reallocated_Sector_Ct   0x0033   100   100   010    Pre-fail  
>>>>> Always        -       0
>>>>>    7 Seek_Error_Rate         0x000f   070   060   045    Pre-fail  
>>>>> Always        -       10214727
>>>>>    9 Power_On_Hours          0x0032   100   100   000    Old_age   
>>>>> Always        -       726h+57m+29.053s
>>>>>   10 Spin_Retry_Count        0x0013   100   100   097    Pre-fail  
>>>>> Always        -       0
>>>>>   12 Power_Cycle_Count       0x0032   100   100   020    Old_age   
>>>>> Always        -       63
>>>>> 183 Runtime_Bad_Block       0x0032   100   100   000    Old_age   
>>>>> Always        -       0
>>>>> 184 End-to-End_Error        0x0032   100   100   099    Old_age   
>>>>> Always        -       0
>>>>> 187 Reported_Uncorrect      0x0032   100   100   000    Old_age   
>>>>> Always        -       0
>>>>> 188 Command_Timeout         0x0032   100   100   000    Old_age   
>>>>> Always        -       0 0 2
>>>>> 189 High_Fly_Writes         0x003a   100   100   000    Old_age   
>>>>> Always        -       0
>>>>> 190 Airflow_Temperature_Cel 0x0022   060   044   040    Old_age   
>>>>> Always        -       40 (Min/Max 39/40)
>>>>> 191 G-Sense_Error_Rate      0x0032   100   100   000    Old_age   
>>>>> Always        -       0
>>>>> 192 Power-Off_Retract_Count 0x0032   100   100   000    Old_age   
>>>>> Always        -       94
>>>>> 193 Load_Cycle_Count        0x0032   100   100   000    Old_age   
>>>>> Always        -       315
>>>>> 194 Temperature_Celsius     0x0022   040   056   000    Old_age   
>>>>> Always        -       40 (0 16 0 0 0)
>>>>> 195 Hardware_ECC_Recovered  0x001a   073   064   000    Old_age   
>>>>> Always        -       20195647
>>>>> 197 Current_Pending_Sector  0x0012   100   100   000    Old_age   
>>>>> Always        -       0
>>>>> 198 Offline_Uncorrectable   0x0010   100   100   000    Old_age   
>>>>> Offline      -       0
>>>>> 199 UDMA_CRC_Error_Count    0x003e   200   200   000    Old_age   
>>>>> Always        -       0
>>>>> 240 Head_Flying_Hours       0x0000   100   253   000    Old_age   
>>>>> Offline      -       237h+57m+03.832s
>>>>> 241 Total_LBAs_Written      0x0000   100   253   000    Old_age   
>>>>> Offline      -       8646277300
>>>>> 242 Total_LBAs_Read         0x0000   100   253   000    Old_age   
>>>>> Offline      -       888924720
>> 
>> SMART data is not the most reliable to tell anything about anything, 
>> but, are the Raw_Read_Error_Rate, Seek_Error_Rate and 
>> Hardware_ECC_Recovered number normal for an SMR drive? It's been ages 
>> since I last had a spinning rust attached, so, I picked up one of my 
>> old Seagate Barracudas..
>> 
>>   1 Raw_Read_Error_Rate     0x000f   101   087   006    Pre-fail  
>> Always       -       3467552
>>   7 Seek_Error_Rate         0x000f   089   060   030    Pre-fail  
>> Always       -       809092981
>>   9 Power_On_Hours          0x0032   044   044   000    Old_age   
>> Always       -       49295
>> 195 Hardware_ECC_Recovered  0x001a   021   003   000    Old_age   
>> Always       -       3467552
>> 
>> To my eyes, it seems like your drive isn't too happy about something. 
>> I've seen disks become unhappy due to excess vibration.. Combining 
>> vibration, SMR and consumer drive might cause the firmware to reset 
>> the drive instead of timing out?
> 
> I wondered about that as well. None of my other non-SMR drives have 
> anything close to those numbers. Unfortunately I have no idea how to 
> tell if that's normal or how to diagnose that as the issue. The drive 
> is laying next to the server right now so the only vibration it should 
> be experiencing is its own and it's still randomly dropping off the 
> controller. At least that what I think is happening based on the 
> "detached" message in dmesg. It is a drive that was in an external USB 
> enclosure and I extracted it. I wonder if Seagate did some nasty "reset 
> randomly if not in the enclosure" thing, I really hope not.
> 
> --
> Jonathan

Seagate are stupid, but not Apple-level of stupid?

Have you tried plotting all the data you can get about your storage 
system (netdata might be the simplest thing to set up to see all the 
things over time), to see if there's a corelation to something?

I would've tossed the drive as soon as it starts causing problems. A 
drive dropping offline (and, hence, triggering higher load on other 
drives when it needs to be resilvered) isn't really helping your array. 
It might do better as (one of more than one) target drive for backups, 
rather than having it in an active array..

Just my €0.2