Banana Pi loosing Harddrive (Sata)

[fightforlife]

Hey there,

 

I am a long time user of your armbian operating system. It is great, I am using it for a simple home NAS system.

 

But I have found one problem with armbian, which is very annoying:

My standard Banana Pi keeps loosing the HDD Ive connected over Sata. It is a 3,5" Seagate Barracude 3TB drive formated as Ext4. It gets 5V and 12V power and runs quite perfect, as long as it has something to do. It spins up and down as I configured it.

But when I leave the pi for a longer time period unattended (no workload), and the harddrive stays in spindown-mode too long, it wont identify the HDD anymore. The mount point and the /dev/sda are gone.

I need to powercycle the whole system to get everything back up.

 

Things I tried:

fresh armbian, with fresh formated drive -> same problem

fresh armbian, with another drive (WD green) -> same problem

 

I noticed this problem quite a while a go, but never got to posting it. Hope somebody has an answer.

Thanks in advance

[arox]

I suppose your system is on SD card and you simply mount the HDD via fstab ?

 

I use a banana PI and a HDD with external power without that sort of problem (except temp and power supply). But I never tried to spin down the disk. Maybe I could try if you dont get bug info about that.

[arox]

I suppose your system is on SD card and you simply mount the HDD via fstab ?

 

Hum ! You said the mount point disappeared ? Are you using some sort of automounter ?

[zador.blood.stained]

@fightforlife

 

Please provide dmesg log (output of "dmesg" command in terminal) or output of "sudo armbianmonitor -u" after device node disappears while HDD is still connected

[fightforlife]

Thanks for all your answers.

 

Yes, the system is installed on a Sandisk Ultra card.

I mount simply by using the mount command. No automounter script or something.

 

 

Here is the armbianmonitor output. It is not directly after the disk disappears, but i think there is a lot going on in the last bit with ata1 / SDA / Sata / Ext4.  

I guess the problem starts at around 6652.232267. 

http://sprunge.us/FCiR

[arox]

Thanks for all your answers.

 

Yes, the system is installed on a Sandisk Ultra card.

I mount simply by using the mount command. No automounter script or something.

 

 

Here is the armbianmonitor output. It is not directly after the disk disappears, but i think there is a lot going on in the last bit with ata1 / SDA / Sata / Ext4.  

I guess the problem starts at around 6652.232267. 

http://sprunge.us/FCiR

I also have ata link reset but dont find failed reset as in your logs :

 

[ 6707.788508] ata1: hard resetting link

[ 6710.006605] ata1: SATA link down (SStatus 1 SControl 310)

[ 6715.016886] ata1: hard resetting link

[ 6717.238422] ata1: SATA link down (SStatus 1 SControl 310)

[ 6722.247381] ata1: hard resetting link

[ 6724.469028] ata1: SATA link down (SStatus 1 SControl 310)

 

And the no "ata1.00: disabled".

 

You should perhaps disable spin down to verify if it is really related.

[tkaiser]

Did you check the HDD with 'smartctl -a /dev/sda' (SMART attribute 199 is the interesting one)?

[fightforlife]

I got a new error: 

[ 1241.455061] ata1: exception Emask 0x10 SAct 0x0 SErr 0x4000000 action 0xe frozen
[ 1241.455087] ata1: irq_stat 0x00000040, connection status changed
[ 1241.455108] ata1: SError: { DevExch }
[ 1241.455153] ata1: limiting SATA link speed to 1.5 Gbps
[ 1241.455178] ata1: hard resetting link
[ 1243.658265] ata1: SATA link down (SStatus 1 SControl 310)
[ 1243.658325] ata1: EH complete

 

Log: http://sprunge.us/aJTh

 

 

 

 

Smartctl output: http://sprunge.us/cdWO

199 UDMA_CRC_Error_Count 0x0032 200 200 000 Old_age Always - 10

 

 

I disabled all the Smart energy saving stuff like APM, Spindown and Standby. I will see if the problem persists.

fresh boot dmesg log: http://sprunge.us/CaDe

[tkaiser]

Smartctl output: http://sprunge.us/cdWO

199 UDMA_CRC_Error_Count 0x0032 200 200 000 Old_age Always - 10

 

 

Well, CRC errors for the Barracuda would've been more interesting but anyway: the error messages might relate to a 'simple' hardware problem, I would check cable and contacts (BTW: the average SATA port and cable as used with Bananas are rated for 50 matings max -- the eSATA connector is specified for 5,000 matings instead)

[fightforlife]

I have used both drives in my PC using Windows 10 without problems (formated as NTFS)

 

Here is the Smart output of the Seagate drive currently in my PC:

[C7] UltraDMA/SATA CRC Error Rate:    200/Always OK, Worst: 200

 

 

I will check cables but I didnt use them that much, 98% of the time the pi is sitting there without being touched. 

 

Also around two month ago I tried the simpleNas distribution with integrated openmediavault for around 30 days. And there the drive never got lost, even with all smart power saving stuff enabled.

 

I will see if the drive gets lost without power saving.

[RagnerBG]

I got a new error: 

[ 1241.455061] ata1: exception Emask 0x10 SAct 0x0 SErr 0x4000000 action 0xe frozen

[ 1241.455087] ata1: irq_stat 0x00000040, connection status changed

[ 1241.455108] ata1: SError: { DevExch }

[ 1241.455153] ata1: limiting SATA link speed to 1.5 Gbps

[ 1241.455178] ata1: hard resetting link

[ 1243.658265] ata1: SATA link down (SStatus 1 SControl 310)

[ 1243.658325] ata1: EH complete

I have similar issue with SATA disk in LamoboR1 and OpenWrt and very, very bad harddrive with a lot of bad blocks (thousands), but still usable:

[   22.784886] ata1: limiting SATA link speed to 1.5 Gbps
[   22.790050] ata1.00: exception Emask 0x10 SAct 0x2 SErr 0x280100 action 0x6 frozen
[   22.797662] ata1.00: irq_stat 0x08000000, interface fatal error
[   22.803607] ata1.00: cmd 60/00:08:81:41:c8/01:00:17:00:00/40 tag 1 ncq 131072 in
         res 40/00:08:81:41:c8/00:00:17:00:00/40 Emask 0x10 (ATA bus error)
[   22.819024] ata1: hard resetting link
[   23.224812] ata1: SATA link up 1.5 Gbps (SStatus 113 SControl 310)
[   23.278811] ata1.00: configured for UDMA/100
[   23.283211] ata1: EH complete

As you can see my harddrive is working no mater the errors. But i am familiar with the origin of errors and it's self-made sata cable which is a little displaced and i am too lazy to open and fix it. So you better follow the advice above and check the cables and connectivity and maybe power source.

[technik007_cz]

I had couple of crc errors with similar logs few days ago on my BPi with external esata drive connected. Solution? I changed sata cable. Old one went to bin straight...

[technik007_cz]

And the I picked it out from bin, marked "bad" and keeping for testing purposes.

[choss38]

Hi !

 

I had this kind of problem a couple of month ago: the problem was solved when my power source shut definitely down !

With my new power source, I never had the problem so I conclude that my old power source had sometimes some voltage drop !