15.1 and External USB disk docking station

[mdi]

Hi Jocko, recently, I can just get that result (disks that cannot wake up) just letting the drive unused for 15m. Checking dmesg, you will see that it starts giving errors, then it is disconnected.

[fvdw]

currently trials are running with an new kernel 3.13.6 maybe this can solve this issue. It seems to have a somewhat different setup of usb ports.
Jocko already has this kernel and could test it.

[Jocko]

Currently, I tested only with kernel version 3.9.5.

So next days, I will try it but at least the dmesg seems clean after plugging the docking station

Code: Select all

[14129.417314] usb 1-1: new high-speed USB device number 2 using orion-ehci
[14129.568619] usb 1-1: New USB device found, idVendor=152d, idProduct=2551
[14129.575298] usb 1-1: New USB device strings: Mfr=1, Product=11, SerialNumber=3
[14129.582517] usb 1-1: Product: USB Mass Storage
[14129.586949] usb 1-1: Manufacturer: JMicron
[14129.591051] usb 1-1: SerialNumber: 000000000001
[14129.978477] usb-storage 1-1:1.0: USB Mass Storage device detected
[14129.998672] scsi2 : usb-storage 1-1:1.0
[14135.020458] scsi 2:0:0:0: Direct-Access     HDD                       1234 PQ: 0 ANSI: 2 CCS
[14135.030188] fvdw1 disk index before get_sd_name = 1
[14135.035053] No LaCie disk map available
[14135.038930] fvdw2 disk index after get_sd_name = 1
[14135.043701] fvdw3 prefix[0]=s  prefix[1]=d
[14135.047817] fvdw4 diskname = sdb
[14135.052188] sd 2:0:0:0: [sdb] 1953525168 512-byte logical blocks: (1.00 TB/931 GiB)
[14135.060574] scsi 2:0:0:1: Direct-Access     HDD                       1234 PQ: 0 ANSI: 2 CCS
[14135.069522] fvdw1 disk index before get_sd_name = 2
[14135.074455] sd 2:0:0:0: [sdb] Write Protect is off
[14135.079260] No LaCie disk map available
[14135.083082] fvdw2 disk index after get_sd_name = 2
[14135.087878] sd 2:0:0:0: [sdb] Mode Sense: 28 00 00 00
[14135.087948] fvdw3 prefix[0]=s  prefix[1]=d
[14135.092065] fvdw4 diskname = sdc
[14135.095792] sd 2:0:0:0: [sdb] No Caching mode page found
[14135.101789] sd 2:0:0:1: [sdc] 156301488 512-byte logical blocks: (80.0 GB/74.5 GiB)
[14135.109625] sd 2:0:0:0: [sdb] Assuming drive cache: write through
[14135.117029] sd 2:0:0:1: [sdc] Write Protect is off
[14135.121893] sd 2:0:0:1: [sdc] Mode Sense: 28 00 00 00
[14135.122918] sd 2:0:0:1: [sdc] No Caching mode page found
[14135.128326] sd 2:0:0:1: [sdc] Assuming drive cache: write through
[14135.135654] sd 2:0:0:0: [sdb] No Caching mode page found
[14135.141018] sd 2:0:0:0: [sdb] Assuming drive cache: write through
[14135.174155] sd 2:0:0:1: [sdc] No Caching mode page found
[14135.179513] sd 2:0:0:1: [sdc] Assuming drive cache: write through
[14135.233407]  sdb: sdb1 sdb2 sdb3 sdb4 < sdb5 sdb6 sdb7 sdb8 >
[14135.269537]  sdc: sdc1 sdc2 sdc3 sdc4 < sdc5 sdc6 sdc7 sdc8 >
[14135.278536] sd 2:0:0:0: [sdb] No Caching mode page found
[14135.283840] sd 2:0:0:0: [sdb] Assuming drive cache: write through
[14135.290007] sd 2:0:0:0: [sdb] Attached SCSI disk
[14135.296771] sd 2:0:0:1: [sdc] No Caching mode page found
[14135.302131] sd 2:0:0:1: [sdc] Assuming drive cache: write through
[14135.308233] sd 2:0:0:1: [sdc] Attached SCSI disk
[14150.016672] EXT3-fs (sdc2): warning: maximal mount count reached, running e2fsck is recommended
[14150.025424] kjournald starting.  Commit interval 5 seconds
[14150.038675] EXT3-fs (sdc2): using internal journal
[14150.044036] EXT3-fs (sdc2): recovery complete
[14150.048415] EXT3-fs (sdc2): mounted filesystem with ordered data mode
[14153.975796] EXT3-fs (sdb7): warning: maximal mount count reached, running e2fsck is recommended
[14153.984575] kjournald starting.  Commit interval 5 seconds
[14153.990920] EXT3-fs (sdb7): using internal journal
[14153.996535] EXT3-fs (sdb7): recovery complete
[14154.000927] EXT3-fs (sdb7): mounted filesystem with ordered data mode
[14157.076297] EXT3-fs (sdc1): warning: maximal mount count reached, running e2fsck is recommended
[14157.085046] kjournald starting.  Commit interval 5 seconds
[14157.099295] EXT3-fs (sdc1): using internal journal
[14157.104658] EXT3-fs (sdc1): recovery complete
[14157.109028] EXT3-fs (sdc1): mounted filesystem with ordered data mode
[14160.125668] EXT3-fs (sdb1): warning: checktime reached, running e2fsck is recommended
[14160.133607] kjournald starting.  Commit interval 5 seconds
[14160.139795] EXT3-fs (sdb1): using internal journal
[14160.145160] EXT3-fs (sdb1): recovery complete
[14160.149551] EXT3-fs (sdb1): mounted filesystem with ordered data mode
[14170.229202] EXT3-fs (sdb5): warning: maximal mount count reached, running e2fsck is recommended
[14170.237945] kjournald starting.  Commit interval 5 seconds
[14170.243918] EXT3-fs (sdb5): using internal journal
[14170.249159] EXT3-fs (sdb5): recovery complete
[14170.253501] EXT3-fs (sdb5): mounted filesystem with ordered data mode
[14173.371650] EXT3-fs (sdc8): warning: maximal mount count reached, running e2fsck is recommended
[14173.380376] kjournald starting.  Commit interval 5 seconds
[14173.386672] EXT3-fs (sdc8): using internal journal
[14173.404107] EXT3-fs (sdc8): recovery complete
[14173.408536] EXT3-fs (sdc8): mounted filesystem with ordered data mode


[mdi]

Hi guys, two considerations, sad ones:
- within Linux, plugging and unplugging the disks works ok. I only got the issue once the disks were plugged and after they timeout. So, after plugging, we need to wait for them to timeout and spin down.
- today I'm testing under Windows and the disk were able to properly wakeup instead :sob

So, under windows (with no special driver loaded), it seems to work, but not under Linux. Very bad news....

[fvdw]

this is a very strange issue, if it works under windows, it should also work under Linux.
In principle the disk doesn't know that it is conected to a sata usb interface and to which operating system the usb hub is connected. it should wake up whenever a read request is made. The question is now why does this read request that works with at least 3 diks I have here and not with your disks :scratch

What happens if you connect a single disk to the usb port ?

[Jocko]

Hi mdi,

I begin to understand nothing with this device. :sob

For the last 24 hours, the two external disks stay spin up all of time (the internal disk spins down) :shock: :shock: :shock:
but a good new, the disks aren't unmounted (no additional lines in dmesg after NAS booting) and data stay available. :-D

So I go to reinstall the previous kernel version, to see if I got the same result. :scratch

[mdi]

Ok Jocko, please check that out with the previous Kernel. I really hope it helps :please

fvdw, yes, you're right, even if I'm not sure if the disks manage its own wake up when a r/w request is make or if the SO sends some wake-up SATA command knowing that the disk spun-down.
I don't have much knowledge on these disks, but everybody says that through an USB connection the SATA communication can be a bit more limited that if the disks are mounted internally in the PC.

So I really don't know, but sure the USB connection is different (I cannot understand why myself, but it's true that I didn't checked that out).

Thanks for considering the issue though!

[Jocko]

After more 1h30, I have the same behavior with the kernel version 3.9.5 :
* 2 USB disks still spin up !!!
* not unmounted and data still available

[fvdw]

@mdi

there are two daemons running that have influence on hdd, these are
(1) noflushd. this one is used to prevent that the internal harddisk spins up by caching some write info.
never seen that it had influence on usb disk but you could try to kill it and see if then the problem is solved.

Code: Select all

killall -9 noflushd

you can check if it is gone by looking in the process list (ps -ef)

(2) syncawayd
This is a daemon that was present already in the original firmware when we start to develop it, I never have been able to figure out what it is actually doing (no source code available) so we left it in peace. I suspect that it syncs data to disk to prevent in case of poer failure that data is lost that is still in cache. But I alos read that it has to do something with syncing between nas and usb..
Also kill this one I would suggest as trial.

[mdi]

Hi guys, i will try killing the daemons.

The curious fact is that i have the same issue when connecting those disks (through a usb docking station, add jocko knows) also to others Linux boxes. I tried with a RPi and with an Ubuntu virtual machine running under Windows.

I also tried with one disk at a time, and one or two different disks. I could try with seagate disks and wd red disks (the ones i would like to use).

Today i will check further with Windows to be sure about differences with the so.
i will let you know.