freeze & reboot issue

[nasos]

We recently built a new FreeNAS server to replace our aging one (also FreeNAS). After moving all the data to the new server and making it the primary storage server in our network we started experiencing the following symptoms:

It freezes up at random intervals (a few times per day)
It sometimes reboots on its own accord



The specs of the new machine are:

Freenas v8.3.1 x64
Motherboard: ASRock x79 Extreme9
CPU: Intel Core i7-3820 3.60GHz
RAM: 64GB (8 x 8GB)
HD: 32TB 8 x 4Tb raidZ (24 TB Available)
GPU: Nvidia geforce 7200GS
PSU: Thermaltake Toughpower 1500W
10G:LAN card.



Any help in troubleshooting this issue will be greatly appreciated. Thank you.

 

[cyberjock]

I'd start with a RAM test to be honest. If the system is freezing there's not much you're going to get from logs.

How long has this system been running? Could you have a hardware incompatibility?

 

[nasos]

Thanks for your help.

The system has been running for almost a month now. But 2/3 of that time we have been rsyncing it with the old server. Everything was working out just fine for five days after the change, and then the malfunctions began. The hardware are commpatible, and during the rsync process we encounered no problems.

 

[pmcnano]

I just registered to the forum, I just built a freenas server for a customer and im experiencing the freezes. I have noticed something weird though. When it freezes, theres no ping, no respone on console, or anything, but the weirdest is that the "reset" button of the computer doesnt actually reset the computer. I have to power down and turn on again (yes the reset button works perfectly fine lol).

I built an Asus H77 Motherboard
i3 2120 3.3Ghz
8GB Ram
4x 3TB HDD.

 

[jgreco]

That'd be a hardware problem, then, wouldn't it. Really, a consumer-grade motherboard doesn't belong in a device you build for a customer. FreeNAS and FreeBSD are likely to stress your hardware in ways that a Windows install won't, and since most ASUS boards include Realtek ethernets, which are a very poor choice for a server, there's a lot of nonoptimal going on here.

 

[pmcnano]

I did a memtest and it passed 6x. The use for the "nas" is pretty basic. Nothing really stressing in my opinion. Though Im new to this. Do you think the Ethernet card could cause trouble? Should I buy an Intel card?

Also @OP sorry for intruding. Should I make a new post?

edit:

Using a 500Watts Cooler Master PSU. Also, this is day 1 using it.

 

[jgreco]

Your opinion of what is stressy is irrelevant. Servers are stressy environments because they are pushing the hardware; you will not find a desktop PC running Windows to be constantly simultaneously shoveling bits to and from disk and network, moving gigabytes per second to and from memory, processing it all on multiple cores, exercising every bit of the I/O subsystem, requiring everything to be organized and to move with exacting precision. It actually used to be a bit worse, because we had to deal with things like shared SCSI busses...

Anything and everything could cause trouble. If you're using a Realtek ethernet, that's bad. If you've picked a non-ECC CPU, that's bad. If you've picked non-ECC memory, that's bad. If you've picked a non-server-grade motherboard, that's bad. Because all those components are designed to work together to form a coherent platform to robustly move around data without the issues people can and do run into with chump-grade hardware from the local PC shop. If you're building hardware for customers and you cannot be bothered to qualify the components before selling them, that's bad. Down that path lies even more badness.

So here's the thing. You can sit there and play exchange-the-parts trying to isolate which specific thing is causing problems. That will maybe fix the freeze issue. It doesn't address the more fundamental problem ... your customer is relying on you to provide good advice and to sell them a gadget that stores and protects their data. What you have described isn't that.

But when even the reset button doesn't work, that's not FreeNAS. That's hardware. And when something fails that catastrophically, that's pretty bad.

 

[cyberjock]

I deleted the previous post from the OP(if you are reading this its already deleted). It had inappropriate language and added no value to the thread except for him/her to vent his frustration. As such I am closing this thread since there is no further valid discussion regarding the topic.