Getting the following message: Watchdog timeout -- resetting

[HHawk]

Hi guys,

I am in need of some help.
It happens not all the time, but from time to time when I watch a movie or something, I notice it loses the connection. Then it takes some time (about 20 or 30 seconds) before the connection is back again. Like stated above it does not happen every time.

Things to note; I recently replaced my single ethernet network cards to an Intel Quad Port network car (LACP), along with a single PCI Intel network card (for Wake-on-Lan & Wake-on-Wan functions).
The switch I have is LACP capable of course.

As far as I can tell, the problem is caused (as shown below) by em4. The card which is em4 is the PCI-based Intel network card.
...I really do not understand why it's even bothering with this card, because the connection should go through the lagg0 (based on em0, em1, em2 and em3) anyways and not through em4.

Since it should run through lagg0, the problems should do anything to the current connection, right?

Jan 15 23:16:47 freenas kernel: em4: Watchdog timeout -- resetting
Jan 15 23:16:47 freenas kernel: em4: link state changed to DOWN
Jan 15 23:16:50 freenas kernel: em4: link state changed to UP
Jan 15 23:17:06 freenas INADYN[3039]: INADYN:IP: Error '0x0' resolving host name 'checkip.dyndns.org'
Jan 15 23:17:06 freenas INADYN[3039]: W: DYNDNS: Error 'RC_IP_INVALID_REMOTE_ADDR' (0x12) when talking to IP server
Jan 15 23:17:06 freenas INADYN[3039]: W:'RC_IP_INVALID_REMOTE_ADDR' (0x12) updating the IPs. (it 54)
Jan 15 23:18:28 freenas INADYN[3039]: INADYN:IP: Error '0x0' resolving host name 'checkip.dyndns.org'
Jan 15 23:18:28 freenas INADYN[3039]: W: DYNDNS: Error 'RC_IP_INVALID_REMOTE_ADDR' (0x12) when talking to IP server
Jan 15 23:18:28 freenas INADYN[3039]: W:'RC_IP_INVALID_REMOTE_ADDR' (0x12) updating the IPs. (it 54)
Jan 15 23:18:43 freenas kernel: em4: Watchdog timeout -- resetting
Jan 15 23:18:43 freenas kernel: em4: link state changed to DOWN
Jan 15 23:18:46 freenas kernel: em4: link state changed to UP

If anyone knows how to fix this, or what to do, please advice.

Thank you very much in advance.

 

[cyberjock]

What version of FreeNAS are you using? Also what is the environment(is this a home server)?

 

[Stephens]

http://www.lmgtfy.com/?q=FreeNAS+watchdog+timer

 

[HHawk]

http://www.lmgtfy.com/?q=FreeNAS+watchdog+timer

Ehr you didn't think I tried searching with Google before? I already did that. Even looked up FreeBSD + Watchdog. Which didn't help me at all!
And if you read my post, you would have known that changing the NIC's didn't help me at all (which is the first result in Google).

People do anything nowadays to get their post count up, sheesh.


@noosauce80

Latest version, non beta. Yes, running it at home.

 

[cyberjock]

In a home environment LACP is useless(and if not properly setup can cause problems too). I learned this the hard way and its mentioned all over the forum that LACP only works when you have LOTS of workstations making LOTS of connections to the server simultaneously. So unless you have more than 20 or so machines I'd say you are not using KISS(Keep it Simple Stupid). I did the same thing 3 years ago so don't feel bad.

So I'd try getting rid of that for starters and see if the problem goes away. If I were a betting man I'd bet $50 that if you got rid of LACP/lagg0 the problem will be gone :)

 

[HHawk]

Useless? I have 5 PC's connected to it and 2 laptops. On 2 of the PC's I work with video editing software which consumes a lot of bandwith.
I also have my AV-receiver connected to it which streams music.

So really, LACP is not ALWAYS useless!

Secondly; if you read my previous post correctly, you should have read and seen that em4 is causing the problems!
But no problem, I will repeat myself once again. The LACP (lagg0) consists out of em0, em1, em2 and em3 (see NOT em4).

And FYI information, with the previous versions of Freenas, for example 7, I never ever had any problems when running LACP with any of my hardware. TY.

 

[HHawk]

I understood that LACP, was balancing the connections, rather than limiting it (as you described it).

I really do not agree with what you are saying, almost all PC's (not including the laptops of course) in my house have dual port NIC's.
And if I would do what you told me, I would be limiting each machine to 1 Gb. Which is of course not an option.

I just ran a few iperf sessions from several machines in the house and I achieved easily a bandwith of 3.2 Gbps (for lagg0) without much effort.

Anyways, thank you for your time, however as seeing this is not going to solve anything, I will check my configuration again, because I think the problem is with em4 (the PCI NIC card) and not with the lagg0.

 

[cyberjock]

Guess what? I just ran a few iperf sessions from several machines in my house and I got 3.4Gbps(roughly the same as yours) without much effort. You cannot exceed a link speed of 1Gb/sec for any one machine, and you are potentially bottlenecked by using LACP, but whatever. I don't know how to explain it any better.

I'll just say what I said before. You should try going without the LACP and see if the error goes away. :) If it does then you'll be doing a setup like what I explained. And you'll have faster machine-to-server performance where you need it. ;)

You might not be bottlenecked today, but you could be tomorrow, but not Saturday. That's why LACP isn't a good choice.

 

[paleoN]

I am in need of some help.

I would hate for you to supply your network config when trying to diagnose a networking problem. It might have even staved off a bunch of don't use LACP posts or at least reduced then to just one.

Since it should run through lagg0, the problems should do anything to the current connection, right?

Based on your description it sounds like you have your networking misconfigured among other things, em4 dropouts. Of course it's hard to tell with what's been supplied.

Code:

ifconfig -a

netstat -n

Feel free to include any other relevant information and don't be shy about your switch configuration either.