Hi,

 

I'm quite lost. i have recently found that my Unraid build is rebooting when my discs spindown, by hand or using disk settings.
syslog atteched, but i didn't see anything that could explain the reboot.

is it possible that my PSU, which is 10 year old, may cause this? is there any test i can make to check what could cause the problem?

syslog

You should mirror the syslog to the flash drive

 

https://forums.unraid.net/topic/46802-faq-for-unraid-v6/page/2/?tab=comments#comment-781601

 

Then replicate the problem.  After a reboot, post the resulting syslog (logs folder on the flash drive) and grab a diagnostics (Tools - Diagnostics) and post it here also.

Hi,

files attached as requested.
Thanks for your help.

nas-diagnostics-20200511-1500.zip syslog

I don't know if this information is relevant or not, but even if all disks are not spindown at the same time, Unraid will reboot.
Last reboot occure with last disk spin down 8min after the previous one.

Just to test, I have set one parity disk to never spin down while other disks are set to spindown after 1h of inactivity.
Then I haven't experienced any reboot for 16 hours.
I do not understand why when all disks spin down unraid reboot and why until one disk still spin up it will not reboot.
Confusing!

It's rather peculiar - your spindown is 0 1 2 3 then 29.

How is your array set up?

2 Ironwolf 6TB as parity et 3 6TB seagate ST6000DM003 for data, got a nvme 256GB as cache also.
Parity are sdd and sdb, data are sde sdc and sdf.
Does it answer to your question?

edit: 29 is parity 2 drive.
others drives seems correctly numbered: parity is 0, disk 1 => 1 disk 2 => disk 3 => 3.

Edited May 12, 20206 yr by milouz

just tried to spindown last disk that was still up, it take roughly 7 minutes, but it rebooted again.

Just made few new tests, it seems that if i spin down the last disk which is still spinning, then, within the next 10 s, 30 s or 60 s i spin up it back, unraid doesn't reboot.
Weird.

up.

This is a very weird one, are you using some kind of ECO power strip or an ECO port of an UPS? Those can sometimes be configured to cut power once it reaches a set low value, though if it was that I would expect the server to shutdown, not reboot, kind of grasping at straws here.

Nop, no UPS, the server is directly plugged to the electrical outlet

Edited May 15, 20206 yr by milouz

Then I would try a different power supply.

but why the PSU would cause reboot only when all disks spin down? weird no?

Yes, but it's still the most logical cause of the problem, it certainly isn't software related.

While you are at it, turn off any CPU overclock and run your RAM at 2133MHz to eliminate overclocking-related instability.

There's a widespread misunderstanding that instability due to overclocking only happens under heavy load.

Heavy load just makes crashes more likely to happen but it is entirely possible for peculiar instability to happen under barely any load at all.

So, I ordered a new PSU, and it seems it was the cause of the problem.

I really do not understand why reboot  was occuring only while all disk spin down and never at any other moment, but since I changed the PSU the problem disapeared.
At least, if this thread help somone else later, it'll be great.

Thanks all for your help.

3 hours ago, milouz said:

So, I ordered a new PSU, and it seems it was the cause of the problem.

I really do not understand why reboot  was occuring only while all disk spin down and never at any other moment, but since I changed the PSU the problem disapeared.
At least, if this thread help somone else later, it'll be great.

Thanks all for your help.

It was probably the power draw of the spin-up of all the drives after the spin-down that was the cause of the problem if the power supply was not up to it.

I never saw any spin up in the log, and I was taking care to not make any action in unraid and plugins were disabled to avoid any activity.

21 minutes ago, milouz said:

I never saw any spin up in the log, and I was taking care to not make any action in unraid and plugins were disabled to avoid any activity.

Yes, but if the spin up causes a crash then it may never get around to creating the log entry indicating that.  That seems far more likely than the spin-down causing the crash even though the spin-down is the last entry logged, particularly if changing the power supply has helped.