DNS-323-B1 Alt-F 0.1RC5 Disk spin up issue

[Rysiek M]

I have been using Alt-f for month now, and i have faced following issue.

When system starts it sees disks like /dev/sda, /dev/sdb, /dev/sdc - screen status.jpg. Problem appears when we wake up sleeping drives or they are waken up to SMART check. Disks appear in system as /dev/sdc, /dev/sdd, /dev/sde but only when option "Wake up disk to perform test" is checked, if is not, disks disappear from system. Only reboot helps...

I have attached screens from front end, and logs. 

[João Cardoso]

On Tuesday, 3 January 2017 22:59:59 UTC, Rysiek M wrote:

I have been using Alt-f for month now, and i have faced following issue.

When system starts it sees disks like /dev/sda, /dev/sdb, /dev/sdc - screen status.jpg.

And from the log, at bootup the SMART monitoring reports everything OK:

Jan  3 22:24:49 NAS1 daemon.info smartd[660]: Device: /dev/sda [SAT], opened

Jan  3 22:24:49 NAS1 daemon.info smartd[660]: Device: /dev/sda [SAT], ST2000DM001-1ER164, S/N:Z4Z5Y6VJ, WWN:5-000c50-091f87d10, FW:CC46, 2.00 TB

Jan  3 22:24:49 NAS1 daemon.info smartd[660]: Device: /dev/sda [SAT], found in smartd database: Seagate Barracuda 7200.14 (AF)

Jan  3 22:24:49 NAS1 daemon.info smartd[660]: Device: /dev/sda [SAT], enabled SMART Attribute Autosave.

Jan  3 22:24:49 NAS1 daemon.info smartd[660]: Device: /dev/sda [SAT], is SMART capable. Adding to "monitor" list.

Jan  3 22:24:49 NAS1 daemon.info smartd[660]: Device: /dev/sdb, type changed from 'scsi' to 'sat'

Jan  3 22:24:49 NAS1 daemon.info smartd[660]: Device: /dev/sdb [SAT], opened

Jan  3 22:24:49 NAS1 daemon.info smartd[660]: Device: /dev/sdb [SAT], WDC WD20EFRX-68EUZN0, S/N:WD-WCC4M7RNF321, WWN:5-0014ee-20dd1d80e, FW:82.00A82, 2.00 TB

Jan  3 22:24:50 NAS1 daemon.info smartd[660]: Device: /dev/sdb [SAT], found in smartd database: Western Digital Red

Jan  3 22:24:50 NAS1 daemon.info smartd[660]: Device: /dev/sdb [SAT], enabled SMART Attribute Autosave.

Jan  3 22:24:50 NAS1 daemon.info smartd[660]: Device: /dev/sdb [SAT], is SMART capable. Adding to "monitor" list.

Jan  3 22:24:50 NAS1 daemon.info smartd[660]: Monitoring 2 ATA and 0 SCSI devices

But latter on the sda drive starts to not communicated, then errors appear at the sdb drive and the disk bus is reset. That leads for the drives to reappear after a while, and new drive letters assigned to them. During this reset period, that can take some tens of seconds or a few minutes, the SMART monitoring daemon starts complaining and trying sending emails (have you Setup Mail?)

I can't tell if the issues start because of smartd trying to communicate to the drives while they are at sleep or not.

Start stopping the smartd daemon, Services->System, smart, uncheck the "Boot Enable" and Submit. The smartd should stop. For a clean restart, reboot the box. Does the problems reappear?

If not, you can disable spindown (setting a value of zero) and start the SMART daemon, to see if the issue only appears while the drive(s) are in spindown or not. You can also try to put only one drive in spindown at a time.

Have you setup/changed anything  under Disk->Utilities, PowerSave and spindown? Please take a screen shot and post it.

 

Problem appears when we wake up sleeping drives or they are waken up to SMART check. Disks appear in system as /dev/sdc, /dev/sdd, /dev/sde but only when option "Wake up disk to perform test" is checked, if is not, disks disappear from system. Only reboot helps...

And if smartd is not running and the drives are sleeping? Can you awake them normally without issues?

I'm afraid that I don't have a solution, I'm only trying to isolate the issue origin.

PS-I noticed that you don't have installed the "network correction fix for RC5" (you don't have Alt-F packages installed). You ***HAVE*** to do that!

[Rysiek M]

W dniu środa, 4 stycznia 2017 20:34:10 UTC+1 użytkownik João Cardoso napisał:

On Tuesday, 3 January 2017 22:59:59 UTC, Rysiek M wrote:

I have been using Alt-f for month now, and i have faced following issue.

When system starts it sees disks like /dev/sda, /dev/sdb, /dev/sdc - screen status.jpg.

And from the log, at bootup the SMART monitoring reports everything OK:

Jan  3 22:24:49 NAS1 daemon.info smartd[660]: Device: /dev/sda [SAT], opened

Jan  3 22:24:49 NAS1 daemon.info smartd[660]: Device: /dev/sda [SAT], ST2000DM001-1ER164, S/N:Z4Z5Y6VJ, WWN:5-000c50-091f87d10, FW:CC46, 2.00 TB

Jan  3 22:24:49 NAS1 daemon.info smartd[660]: Device: /dev/sda [SAT], found in smartd database: Seagate Barracuda 7200.14 (AF)

Jan  3 22:24:49 NAS1 daemon.info smartd[660]: Device: /dev/sda [SAT], enabled SMART Attribute Autosave.

Jan  3 22:24:49 NAS1 daemon.info smartd[660]: Device: /dev/sda [SAT], is SMART capable. Adding to "monitor" list.

Jan  3 22:24:49 NAS1 daemon.info smartd[660]: Device: /dev/sdb, type changed from 'scsi' to 'sat'

Jan  3 22:24:49 NAS1 daemon.info smartd[660]: Device: /dev/sdb [SAT], opened

Jan  3 22:24:49 NAS1 daemon.info smartd[660]: Device: /dev/sdb [SAT], WDC WD20EFRX-68EUZN0, S/N:WD-WCC4M7RNF321, WWN:5-0014ee-20dd1d80e, FW:82.00A82, 2.00 TB

Jan  3 22:24:50 NAS1 daemon.info smartd[660]: Device: /dev/sdb [SAT], found in smartd database: Western Digital Red

Jan  3 22:24:50 NAS1 daemon.info smartd[660]: Device: /dev/sdb [SAT], enabled SMART Attribute Autosave.

Jan  3 22:24:50 NAS1 daemon.info smartd[660]: Device: /dev/sdb [SAT], is SMART capable. Adding to "monitor" list.

Jan  3 22:24:50 NAS1 daemon.info smartd[660]: Monitoring 2 ATA and 0 SCSI devices

But latter on the sda drive starts to not communicated, then errors appear at the sdb drive and the disk bus is reset. That leads for the drives to reappear after a while, and new drive letters assigned to them. During this reset period, that can take some tens of seconds or a few minutes, the SMART monitoring daemon starts complaining and trying sending emails (have you Setup Mail?)

Nope, i don't use email monitoring. 

 

I can't tell if the issues start because of smartd trying to communicate to the drives while they are at sleep or not.

Start stopping the smartd daemon, Services->System, smart, uncheck the "Boot Enable" and Submit. The smartd should stop. For a clean restart, reboot the box. Does the problems reappear?

I have stopped smartd daemon for a while during the day, but Smart is not a problem i think, because few mins ago i have tried to wake up both disks and they disappeared and appeared after few mins.

 

If not, you can disable spindown (setting a value of zero) and start the SMART daemon, to see if the issue only appears while the drive(s) are in spindown or not. You can also try to put only one drive in spindown at a time.

Have you setup/changed anything  under Disk->Utilities, PowerSave and spindown? Please take a screen shot and post it.

For yesterday test i used default values. One hint is when one of drives doesn't sleep, second is able to wake up normally. - Both disks were bought around month ago. 

 

 

Problem appears when we wake up sleeping drives or they are waken up to SMART check. Disks appear in system as /dev/sdc, /dev/sdd, /dev/sde but only when option "Wake up disk to perform test" is checked, if is not, disks disappear from system. Only reboot helps...

And if smartd is not running and the drives are sleeping? Can you awake them normally without issues?

If smartd is running and not running both drives are sleeping, but i cant wake them up manually. Also i have observed that if i press button "sandbynow" and wake them after 5mins, it will be ok. After hour they will disappear.

 

I'm afraid that I don't have a solution, I'm only trying to isolate the issue origin.

PS-I noticed that you don't have installed the "network correction fix for RC5" (you don't have Alt-F packages installed). You ***HAVE*** to do that!

I use user script with this command: /Alt-F/usr/sbin/ethtool -K eth0 tso off

 

 

I have attached screens from front end, and logs. 

For test i have removed one disk, and situation is the same disk went sleep at 22:07CET, i tried to wake it up at 22:47 disk disappeared.

 

[João Cardoso]

On Wednesday, 4 January 2017 20:38:35 UTC, Rysiek M wrote:

W dniu środa, 4 stycznia 2017 20:34:10 UTC+1 użytkownik João Cardoso napisał:

On Tuesday, 3 January 2017 22:59:59 UTC, Rysiek M wrote:

I have been using Alt-f for month now, and i have faced following issue.

When system starts it sees disks like /dev/sda, /dev/sdb, /dev/sdc - screen status.jpg.

And from the log, at bootup the SMART monitoring reports everything OK:

Jan  3 22:24:49 NAS1 daemon.info smartd[660]: Device: /dev/sda [SAT], opened

Jan  3 22:24:49 NAS1 daemon.info smartd[660]: Device: /dev/sda [SAT], ST2000DM001-1ER164, S/N:Z4Z5Y6VJ, WWN:5-000c50-091f87d10, FW:CC46, 2.00 TB

Jan  3 22:24:49 NAS1 daemon.info smartd[660]: Device: /dev/sda [SAT], found in smartd database: Seagate Barracuda 7200.14 (AF)

Jan  3 22:24:49 NAS1 daemon.info smartd[660]: Device: /dev/sda [SAT], enabled SMART Attribute Autosave.

Jan  3 22:24:49 NAS1 daemon.info smartd[660]: Device: /dev/sda [SAT], is SMART capable. Adding to "monitor" list.

Jan  3 22:24:49 NAS1 daemon.info smartd[660]: Device: /dev/sdb, type changed from 'scsi' to 'sat'

Jan  3 22:24:49 NAS1 daemon.info smartd[660]: Device: /dev/sdb [SAT], opened

Jan  3 22:24:49 NAS1 daemon.info smartd[660]: Device: /dev/sdb [SAT], WDC WD20EFRX-68EUZN0, S/N:WD-WCC4M7RNF321, WWN:5-0014ee-20dd1d80e, FW:82.00A82, 2.00 TB

Jan  3 22:24:50 NAS1 daemon.info smartd[660]: Device: /dev/sdb [SAT], found in smartd database: Western Digital Red

Jan  3 22:24:50 NAS1 daemon.info smartd[660]: Device: /dev/sdb [SAT], enabled SMART Attribute Autosave.

Jan  3 22:24:50 NAS1 daemon.info smartd[660]: Device: /dev/sdb [SAT], is SMART capable. Adding to "monitor" list.

Jan  3 22:24:50 NAS1 daemon.info smartd[660]: Monitoring 2 ATA and 0 SCSI devices

But latter on the sda drive starts to not communicated, then errors appear at the sdb drive and the disk bus is reset. That leads for the drives to reappear after a while, and new drive letters assigned to them. During this reset period, that can take some tens of seconds or a few minutes, the SMART monitoring daemon starts complaining and trying sending emails (have you Setup Mail?)

Nope, i don't use email monitoring. 

 

I can't tell if the issues start because of smartd trying to communicate to the drives while they are at sleep or not.

Start stopping the smartd daemon, Services->System, smart, uncheck the "Boot Enable" and Submit. The smartd should stop. For a clean restart, reboot the box. Does the problems reappear?

I have stopped smartd daemon for a while during the day, but Smart is not a problem i think, because few mins ago i have tried to wake up both disks and they disappeared and appeared after few mins.

You have to exclude one thing at a time. While you see in the system log "hard resetting link", you still have the problem. Its better to use the System log because it is time-stamped.

You have already disabled smartd on boot. keep it disabled just to continue excluding it.

Now power down the box and remove the USB pen and one of the drives. Power up, and if the issue reappears, power down, remove the disk and insert the other one. If the issue reappears, then the box itself is the issue.

And the box itself can have an issue in just one of the bays, try identifying it using the same method -- one drive has no issues on the left bays and has on the right bay? that's a faulty bay in the box...

When you think that you have identified the problematic drive or bay, remove it and boot with only the drive that you think it is OK and the issue should not reappear.

Yes, I know, it's boring and lengthy. 

Regarding the "network corruption fix" -- the logs shows that Alt-F packages are not installed. Does Packages->Alt-F shows a list of installed/pre-installed packages?

Anyway as packages are installed on disk, with all that disk insertion and removal you will have to deffer that to a latter time.

[Rysiek M]

It seems that power supply has died.
It gives 5V and 10,5V, as I can see power supply should give 5 and 12. I have exchanged it to different one, luckily I had one with similar plug.

[João Cardoso]

On Saturday, 14 January 2017 20:21:02 UTC, Rysiek M wrote:

It seems that power supply has died.
It gives 5V and 10,5V, as I can see power supply should give 5 and 12. I have exchanged it to different one, luckily I had one with similar plug.

So, is problem solved? I would never thought on the 12V power supply...

[Rysiek M]

Neither do I... It seems that 12V is needed by HD to spin up, then it can work with only 5V.

I think, issue can be marked as solved. Because i tried three times during last 24h, and every time drives were waking up without any issue.