Weird networking problems after 60 days of uptime

[ooimo]

Everytime my NAS gets to 60ish days of uptime the networking stops working and only works again once I reboot.

I'm unable to ping the machine at all, once I disconnect and reconnect the Ethernet I'm then able to ping it but can't ssh in as there is no route to host.

I have no idea how to go about fixing this since there is never any sign of problems until I hit around 60 days of uptime.

I'm using an Intel CT gigabit Ethernet card. I used to use the motherboard's ethernet port until I had this issue a while ago and thought a better card would fix it.

 

[Nick2253]

FreeNAS version? Hardware specs?

 

[DifferentStrokes]

Also supply some general network layout and specs if you can. If you had the problem before with the Mobo NIC and now with the Intel, makes me wonder if the problem is something on the network or cabling or something.... Just kinda thinking out loud here.

 

[ooimo]

FreeNAS version? Hardware specs?

Build: FreeNAS-9.2.1.9-RELEASE-x64 (2bbba09)
Platform: AMD A8-5600K APU with Radeon(tm) HD Graphics
Memory: 7348MB

/var/log/messages had nothing since yesterday and it had been working fine all of today

/var/log/dmesg looks like this:

Code:

[root@freenas] ~# cat /var/log/dmesg.today
arp: 192.168.1.164 moved from 7c:7a:91:2d:fc:71 to 3c:97:0e:eb:4f:a1 on em0
[root@freenas] ~# cat /var/log/dmesg.yesterday
arp: 192.168.1.15 moved from 02:58:26:00:12:0a to 68:05:ca:20:1c:76 on epair8b
arp: 192.168.1.15 moved from 02:77:ca:00:11:0a to 68:05:ca:20:1c:76 on epair7b
arp: 192.168.1.15 moved from 02:d5:16:00:10:0a to 68:05:ca:20:1c:76 on epair6b
arp: 192.168.1.15 moved from 02:cc:68:00:0f:0a to 68:05:ca:20:1c:76 on epair5b
arp: 192.168.1.15 moved from 02:f6:47:00:0e:0a to 68:05:ca:20:1c:76 on epair4b
arp: 192.168.1.15 moved from 02:21:54:00:0d:0a to 68:05:ca:20:1c:76 on epair3b
arp: 192.168.1.15 moved from 02:39:a7:00:0c:0a to 68:05:ca:20:1c:76 on epair2b
arp: 192.168.1.15 moved from 02:91:25:00:0b:0a to 68:05:ca:20:1c:76 on epair1b
arp: 192.168.1.15 moved from 02:52:40:00:0a:0a to 68:05:ca:20:1c:76 on epair0b
[root@freenas] ~#

 

[Nick2253]

I would recommend upgrading to 9.3. At the very least, that will rule out a misc bug that got fixed.

 

[Darren Myers]

Build: FreeNAS-9.2.1.9-RELEASE-x64 (2bbba09)
Platform: AMD A8-5600K APU with Radeon(tm) HD Graphics
Memory: 7348MB

This doesnt really answer the HW question. Such as motherboard make/model,harddrives make/model/size/how many, things like that.

 

[ooimo]

Motherbaord is an Asus A85XM-A. The memory isn't ECC but I've run memtest many times without any errors. I have 5 WD Red drives in a RAIDz1 configuration, all connected to the motherbaord. There have been no changes to the network recently, no new devices or anything. I don't have link aggregation or VLANs or anything like that.

Also, just as last time, it works fine after a reboot. I would like to find the root cause so I don't have to reboot periodically.

 

[Darren Myers]

Motherbaord is an Asus A85XM-A

And the cause of the issue is, *drum roll* Realtek 8111F LAN.

According to Newegg and confirmed by ASUS

 

[ooimo]

So even though it's completely disconnected it's causing problems? Should disabling it in bios solve this?

 

[Darren Myers]

I'm using an Intel CT gigabit Ethernet card. I used to use the motherboard's ethernet port until I had this issue a while ago and thought a better card would fix it.

Ah didnt see that, my mistake

 

[ooimo]

It hasn't got an interface associated with it, however the system lockup problem may be what I'm experiencing. So I'll disable it in case.

http://lists.freebsd.org/pipermail/freebsd-questions/2014-February/256286.html

 

[SweetAndLow]

How do you know the networking is dying? Can you connect a keyboard and mouse when this happens check logs and the filesystem.

 

[Robert Trevellyan]

Does your box get its IP address via DHCP? If so, what is the lease duration?