Build 636 keeps rebooting randomly

[glouk]

Since i upgraded to build 636, my nas4free box keeps rebooting randomly, with no apparent reason...

Since it's embedded, i have no trace of what happened after reboot.

At first I thought it was a problem with the ups driver, so i disabled the ups service, but to no effect

Any suggestion ? is there a way to have nas4free write the log to another place, so that it does not disappear ?

Thanks.

glouk

[elaps]

Glouk,
You could set up syslog in Nas4Free, and run a syslog server somewhere else. Even if you just have a windows PC you can run a windows syslog server. I have one (syslogrd.exe) that came from Draytek with a router, but I am sure there are others out there.
Alan

[raulfg3]

more easy, configure logs to store it on your data disk ( in /mnt/yourpool/logs for example).

you only need configure clog_logdir as descrive in wiki: http://wiki.nas4free.org/doku.php?id=faq:0046

see my NAS for example:

logs.jpg

PD: In my experience, ramdom reboot are usually RAM (not best tunning or bad hardware)

[glouk]

Thank you all,

The problem seems to be hard drives : i used 2 1,5To samsung drives in a zfs stripe, and they seem to be falling down at the same time : smart fails seem to be the cause of the reboots.

If there are new developments, i'll post them here

[RedAntz]

Thank you all,

The problem seems to be hard drives : i used 2 1,5To samsung drives in a zfs stripe, and they seem to be falling down at the same time : smart fails seem to be the cause of the reboots.

If there are new developments, i'll post them here

Hi glouk,

What do you mean by 'smart' fails ?

[glouk]

smart daemon reports that the disk is failing/unavailable...

btw, i'll also turn to a ram tester, since changing the disks did not solve my wild reboots problem...

[rostreich]

Set AHCI to IDE in Bios. Some Samsungs have firmware problems.
There exist some firmware patches, but not for every 1,5er type.

[glouk]

Hi,
I think almost every problem has been solved (changed 2 drives, trashed a dimm, ...)
however, one thing remains : every night at 3:00, my box reboots.
yet, nothing is supposed to happen at that time, i removed a smart check that was scheduled at 3:00, to no effect...
Is it possible that a cron event would exist without showing in the web interface ?

[shakky4711]

Hello,

3.00PM and rebbots, I could remember for this combination and found this thread http://forums.freebsd.org/showthread.php?t=29994

Please do a manual zpool scrub from your pool, then let us know if it runs or resulted in a reboot.

Good luck
Shakky

[glouk]

Thanks for all your help !
In fine, I removed everything planned to happen at or around 3 AM, and found in my configuration a schedule for shutdown at that time every day. I sure never programmed this, so I just deleted this schedule and did not bother to wonder
Everything has been going fine since, except for some AFP restarts that i can't explain, and are not of the taste of the client computers...

[dnar]

Thanks for all your help !
In fine, I removed everything planned to happen at or around 3 AM, and found in my configuration a schedule for shutdown at that time every day. I sure never programmed this, so I just deleted this schedule and did not bother to wonder
Everything has been going fine since, except for some AFP restarts that i can't explain, and are not of the taste of the client computers...

That will do it!!!!

As soon as read "everyday at 3:00am" I thought scheduled reboot...

It's called an ID 10 T error.