SOLVED Network drops out consistently after running for a week or so

[Stingray88]

OK cool, here it is. https://dl.dropboxusercontent.com/u/5826201/debug-freenas-20160225011010..tgz

 

[Robert Trevellyan]

Lifetime Min/Max Temperature: 22/73 Celsius
Lifetime Min/Max Temperature: 22/73 Celsius
Lifetime Min/Max Temperature: 22/72 Celsius
Lifetime Min/Max Temperature: 22/75 Celsius
Lifetime Min/Max Temperature: 22/69 Celsius
Lifetime Min/Max Temperature: 22/72 Celsius

WTF is going on with your drive temperatures? Do you have SMART temp warnings enabled? Does the problem occur during scrubs?

 

[Mirfster]

Hmm, aren't most drives rated at like 60 Celsius max?

 

[Stingray88]

Are those min/max only from the time of the drives in this FreeNAS installation? Or the lifetime of the drive in general? Because these drives were previously in a POS Drobo 5N for about 2 months before it died on me. That could be where those temperatures are from... Because as far as I can tell the drives are averaging in the 30s while in my FreeNAS.

I don't have SMART temp warnings enabled... I assumed they would have been on by default. I'm setting it up now to warn me if it goes above 45C.

 

[Ericloewe]

Hmm, aren't most drives rated at like 60 Celsius max?

Same way their MTBF is supposed to be hundreds of thousands of hours.

In other words, it's bullshit.

 

[Mirfster]

Same way their MTBF is supposed to be hundreds of thousands of hours.

In other words, it's bullshit.

Yeah, but I need to believe the MTBF (mostly) since I buy my drives used and can't change that. At least with the temps, I can control them. ;)

 

[Robert Trevellyan]

the lifetime of the drive in general?

This.

I'm setting it up now to warn me if it goes above 45C.

The consensus here is that over 40C is cause for concern. I have one drive that regularly reaches 41C during scrubs, which I can live with.

as far as I can tell the drives are averaging in the 30s

OK, but the question remains, is there any correlation between scheduled scrubs, or any other regularly scheduled task, and the failures? Don't forget to look for regularly scheduled activity elsewhere on your network too.

 

[Stingray88]

The consensus here is that over 40C is cause for concern. I have one drive that regularly reaches 41C during scrubs, which I can live with.

OK, but the question remains, is there any correlation between scheduled scrubs, or any other regularly scheduled task, and the failures? Don't forget to look for regularly scheduled activity elsewhere on your network too.

Copy that. I've set informational to 40C and Critical to 45C.

The correlation between scheduled tasks is a good question. It shouldn't be scrubs, as I've got that set to go only once every 30 days assuming I set it up properly, and it's dropping network connection a lot sooner than that. I can't think of any other scheduled tasks on FreeNAS beside SMART and scrubs. Do you have any ideas of what I could check?

As far as the rest of the network is concerned, nothing else is really running when I lose connection beside the router and some iOS devices maybe backing up to icloud. All my desktops and laptops are turned off while I'm away. The router is set to renew DHCP on a schedule of course, but FreeNAS is given a static, and that schedule is like once every 48 hours I think... I'm not dropping connection THAT often. More like once a week.

 

[Mirfster]

but FreeNAS is given a static

Is that static an IP outside of your DHCP range?
Do you have a Gateway and Name Server entered (Under [Network] - [Global Configurtaion])?
I know all this is in your debug, but I am too lazy to go through it all. :P

 

[Stingray88]

Is that static an IP outside of your DHCP range?

No.

Do you have a Gateway and Name Server entered (Under [Network] - [Global Configurtaion])?

In FreeNAS or the router? I've never manually set anything like that in either, so do you think that could be an issue?

I must apologize, networking is one of my least knowledgeable subjects in computing. I'm working on it haha.

I know all this is in your debug, but I am too lazy to go through it all. :p

Hey its better to be thorough about this to avoid skipping over the obvious!

 

[Mirfster]

No.

Hmm, well then on your router do you have at least an IP Reservation for your FreeNas (via MAC Address)? If not, then maybe you would rather do that? But as far as if you manually assign an IP (on the FreeNas Server), you want to use an IP that is NOT in your DHCP range.

Example:

• My router is setup to allocate IP addresses (DHCP)
• My DHCP range is 192.168.10.50 - 192.168.10.200
• So on my FreeNas Server, I set the IP to something that is NOT being doled out by DHCP. Say 192.168.10.25...

In FreeNAS or the router? I've never manually set anything like that in either, so do you think that could be an issue?

Was talking about in FreeNas, here is what mine looks like (for reference - "Domain" was redacted):

 

[Stingray88]

Hmm, well then on your router do you have at least an IP Reservation for your FreeNas (via MAC Address)? If not, then maybe you would rather do that? But as far as if you manually assign an IP (on the FreeNas Server), you want to use an IP that is NOT in your DHCP range.

Example:

• My router is setup to allocate IP addresses (DHCP)
• My DHCP range is 192.168.10.50 - 192.168.10.200
• So on my FreeNas Server, I set the IP to something that is NOT being doled out by DHCP. Say 192.168.10.25...

Ah, that's what I meant by a static, that I had reserved an IP on my router for FreeNAS based on the MAC address. Sorry for the confusion...

Was talking about in FreeNas, here is what mine looks like (for reference - "Domain" was redacted:

I just have "freenas" under hostname and "local" under domain. Nothing else is filled in. I'm assuming those are fine to leave, but I don't know what I should put for the Gateway and Name server. Would it be the router IP?

 

[Mirfster]

but I don't know what I should put for the Gateway and Name server. Would it be the router IP?

Yep, that is the way I have mine setup.

 

[Stingray88]

Copy that. I have them set now. I'll report back if the network connection drops again... crossing my fingers its something as simple as that...

 

[Stingray88]

By now I think it's clear having nothing under the default gateway and nameserver was likely my issue. My server has stayed up for two weeks! Never lasted that long without a reboot before. Thanks guys!

 

[Mirfster]

NP, Glad it worked out.

 

[DrKK]

Still, the path to intellectually understanding what happened here, and why this is the solution, requires further ploughing.