SOLVED Random reboots.. mystery

[toyebox]

Hey everyone. First off, i wan't to say thank you for all the hard work everyone here has done to make such a great operating system. I love it. Here are my current specs

core i7-2670QM (mobile processor)
Jetway JNF9G‑QM77 mobo
16 gigs of ddr3 1600 ram (non ecc.. i know. couldn't find a board that supported it and a mobile processor)
4x 4tb seagate drives in a striped mirror

Everything runs amazing. I have a jail with Couchpotato, Sonarr and Plex in it. Then i have a jail with ruTorrent/rtorrent and openvpn installed.

From what i understand cpu temps on mobile processors climb up to 100 degrees celsius. I have only seen as much as 79.

I have been monitoring temps closely because i figured that would be root of the reboots. I cannot pinpoint it rebooting when I'm doing something specific.

Any help would be greatly appreciated. Also, i wrote a script that runs in cron every minute. it is simple; just gets cpu/hd temps and emails them to me if they are above 74 on cpu or 35 on hds. to avoid having it sending me 100 emails.. i added a small timestamp that waits a user set amount of time before sending another. I have tried disabling this, but i highly doubt this is causing issues.

 

[Robert Trevellyan]

Did you burn in all the hardware?

It's possible that motherboard just doesn't play nicely with FreeNAS.

 

[toyebox]

The processor came from a lightly used laptop.
i burned the memory in with memtest86

Are there logs or anything that could indicate why it's restarting?

 

[Robert Trevellyan]

i burned the memory in with memtest86

• Did you burn in the hard drives?
• Have you triple-checked all the cables?
• Do you have a spare PSU you can swap in?

Are there logs or anything that could indicate why it's restarting?

/var/log

 

[toyebox]

the hard drives, no. i have triple checked all the cables. i have already swapped the PSU with a known working one.

Could the hard drives really do this?

 

[Joseph Lennemann]

the hard drives, no. i have triple checked all the cables. i have already swapped the PSU with a known working one.

Could the hard drives really do this?

Yes. I've had a faulty MSI CD burner randomly reboot a Windows machine. It took me over a year to figure it out.

 

[toyebox]

Hmm. could it be a memory thing? here are the main things running on the box. it's spit between two jails, and everything was installed manually instead of through plugins.

plex media server (DLNA disabled)
sonarr
couchpotato
openvpn
rtorrent
rutorrent
sabnzbd
jackett

It looks when i run top that its a lot of python instances (between 3 and 5) which i expected.. but they all range from 150 to 450 MB's or usage.

I am only working with 16gigs of memory which is the max this board will support unfortunately.

I have no options for upgrading. is there anything i can do to mitigate the amount of memory being used? I noticed on average its about 12 Gigs of memory is committed to wired memory. is this pretty much required from ZFS read/write stuff?

I have read some things about adding an SSD for read/write cache. would this help with memory usage? if so, is there information somewhere regarding this?

Thank you so much!

 

[SweetAndLow]

SSD won't help with your workflow. I would burn in your HDD and go from there. After that it could be memory usage from your jails.

Sent from my Nexus 5X using Tapatalk

 

[toyebox]

so i managed to catch some errors before restarting.

Any ideas?

 

[depasseg]

MPSSAS is the disk controller. Looks like there is an issue with that. Have you installed a driver recently, or is this a stock out of the box installation?

 

[toyebox]

it's a fresh out of the box system now. i reinstalled to check that off my list. just happened again 10 minutes after being on. doing absolutely nothing. I am using an m1015 flashed to IT mode.

 

[depasseg]

Given the troubleshooting that you've done so far, if it were me, I'd pull the card out and see if the reboots stop. It's possible you have a faulty or counterfeit card.

 

[toyebox]

I don't want to speak too soon but... I think i have identified the issue as being a super hot chipset on my LSI raid card. through a little fan on it for temporary use to see what it would do.. i couldn't even create a snapshot.

right now i just started a badblocks test on all 4 disks. it was failing within the first minute before, freezing all input. still going strong now. I think it has to do with the airflow in the case.

It definitely looks and acts like a legit card. I felt the heat sink on the chipset and damn, it was pretty hot.

 

[wblock]

The LSI cards are meant to be in airflow, or those heatsinks do get very hot. I've added fans to my converted Dell H200s. Not my idea, got it from here: https://techmattr.wordpress.com/201...-flashing-to-it-mode-dell-perc-h200-and-h310/.

 

[toyebox]

yep.. 9 hours later and everything is still running smooth! it's odd, when I searched Google for the kernel panic I was getting, I couldn't find a thing. I'm glad it's all taken care of. thanks everyone!! on another note, are there real benefits to adding an SSD for read/write cache? I mean.. my speeds are great now lol. I don't know how much faster they can get

 

[SweetAndLow]

With your limited memory a read l2arc will hurt performance. A write slog device only helps with a sync write workflow which you don't do. So no a SSD won't help you at all.

Sent from my Nexus 5X using Tapatalk

 

[toyebox]

Alright. Well thanks again everyone! I am so grateful for the system i have now. an All in one box that saturates a gigabit connection easily lol.

 

[Mirfster]

A write slog device only helps with a sync write workflow which you don't do. So no a SSD won't help you at all.

Ah, but what is one were to set "sync=always"? /Devil's Advocate.... ;)

 

[toyebox]

hehe :P

So for accurate disk performance testing... ideas? i saw bonnie++ but it was kind of confusing. i can run dd with a 100gig file but it comes back with over 4000 mb/s

 

[MrToddsFriends]

i can run dd with a 100gig file but it comes back with over 4000 mb/s

Are you running dd with if=/dev/zero and of= on a dataset with compression turned on? A sequence of zeros is highly compressible.

See also this thread, which keeps going for years.
https://forums.freenas.org/index.php?threads/notes-on-performance-benchmarks-and-cache.981/