Hi,
On Fri, Dec 5, 2014 at 10:12 PM, Nuno Magalhães <nunomagalhaes@xxxxxxxxx> wrote:
> On 2014-12-04 23:59, Phillip Susi wrote:
>> Looks a lot like a drive that has Power Up In Standby enabled; on
>> power up they don't know what to report for many of the fields
>> until the media has been read.
>
> The problem is this only happens sometimes and only on WD drives, but
> i'll look into that, thanks.
The 2 WD disks do have PUIS, but it's disabled by default, so all 4
should power-up on boot (as they mostly do).
Since i've heard the problem could be the PSU, i did some research and
ended up (layman-ly) concluding that a 500W PSU is sufficient for all
4 disks. Still, i wanted to try and enable PUIS on the WDs, maybe it
could work (the PSU isn't new, wouldn't be as stressed on boot, etc).
As far as the Toshibas go, the only tweak they have is enabling TLER.
On rc.local i have:
smartctl -l scterc,70,70 /dev/disk/by-id/ata-TOSHIBA_DT01ACA100_SERIAL01
smartctl -l scterc,70,70 /dev/disk/by-id/ata-TOSHIBA_DT01ACA100_SERIAL02
(These 2 always boot and now hold a RAID1, although lshw says one has
a RAID partition and the other an ext4... but:
# cat /proc/mdstat
Personalities : [raid1]
md0 : active raid1 sda1[0] sdb1[1]
976629568 blocks super 1.2 [2/2] [UU]
unused devices: <none>
...so i'm not (too) worried about lshw.)
As for the WDs...
I went and tried hdparm -s1 --yes-i-know-what-i-am-doing
/dev/disk/by-id/ata-WDC_WD10EFRX-68PJCN0_WD-SERIAL02
It did enable the flag:
* Power Management feature set
* Power-Up In Standby feature set
* SET_FEATURES required to spinup after power up
Device-initiated interface power management
unlike its sibling:
* Power Management feature set
Power-Up In Standby feature set
* SET_FEATURES required to spinup after power up
* Host-initiated interface power management
Device-initiated interface power management
but it still came up on boot.
So i tried hdparm.conf:
/dev/disk/by-id/ata-WDC_WD10EFRX-68PJCN0_WD-SERIAL02 {
poweron_standby = on
}
Booted today and one failed as they usually do, the other hiccuped and
failed also (they don't usually fail simultaneously). I echoed - - -
and they both came up, although the non-PUIS took longer. The
PUIS-enabled one had this on boot:
[ 12.824479] ata4.00: failed to IDENTIFY (SPINUP failed, err_mask=0x4)
which i took as a good sign since i didn't want it spinning up.
As i reinstalled a fresh debian i hadn't enabled ant kernel parameters yet. :(
I'd try another PSU but the ones i have laying around are older. I'd
try enabling PUIS via jumper but i'm not sure the PSU lacking power is
the issue. I guess i could partition and install something onto the
WDs and try and boot them isolated and go from there.
I'm beginning to consider a boot script to count disks as my option,
but i'd rather understand why this is happening rather than hammering
around it.
Attached is info on all drives and the last dmesg.
Any further hints are highly appreciated.
Regards,
Nuno
Attachment:
dmesg
Description: Binary data
Attachment:
SATA01-T1
Description: Binary data
Attachment:
SATA02-W1
Description: Binary data
Attachment:
SATA03-T2
Description: Binary data
Attachment:
SATA04-W2
Description: Binary data
