AppleIntelE1000e Hardware Hang - Ethernet port non-responsive until reboot

[atlantis852]

Cool! Glad I could help.

 

[marklaurel]

Thank you! I was having the same issue with system hangs when transferring large amounts of data through my home network. Rolling back the network driver to v2.4.14 seems to have fixed it.

 

[The LT]

From my testing, it seems that 2.4 driver only manages to partially circumvent the issue. While it's much more stable, it is still possible to crash the NIC, it just requires absolutely full-throttle throughput.

It's easily achievable using iperf or iSCSI throughput, which is more intensive bandwidth-wise than NAS Samba and AFP transfers.

It's worth noting that it happens even with native OS X drivers for Intel NICs so it seems something other than the NIC driver kext is at fault here.

 

[atlantis852]

I make regular heave use of the NIC for production work and haven't been able to reproduce the hang issue under 2.4.14 so far (thankfully). Typically I am transferring 60 to 300 GB sized files at >980 mbps average throughput. These are transfers over the LAN to servers and other machines. I am not running iSCSI so I haven't tested to see if it is a trigger point.

There is definitely something happening to trigger the hardware hang in v3.x as any heavy use can reproduce the issue.

 

[The LT]

I make regular heave use of the NIC for production work and haven't been able to reproduce the hang issue under 2.4.14 so far (thankfully). Typically I am transferring 60 to 300 GB sized files at >980 mbps average throughput. These are transfers over the LAN to servers and other machines. I am not running iSCSI so I haven't tested to see if it is a trigger point.

There is definitely something happening to trigger the hardware hang in v3.x as any heavy use can reproduce the issue.

Let me stress it again, it's happening on native AppleIntel8254XEthernet.kext as well... This can't be hnak's driver issue, it's something else.

I've migrated to Clover from Chameleon and the issue persists.

 

[atlantis852]

Let me stress it again, it's happening on native AppleIntel8254XEthernet.kext as well... This can't be hnak's driver issue, it's something else.

I've migrated to Clover from Chameleon and the issue persists.

Hmmm... Something ugly seems to be going on post 10.9.4 for many people but for me, v2.4.14 seems to have resolved the port hang issue I was having.

I would like like to find the root cause though.

 

[The LT]

Are you positive it started on 10.9.4? I seem to have had issues earlier as well.

Besides: I am a bit heavier loaded than you. I am transferring lots of files of 1gb to 3gb size with total amounts from 400Gb to 2-3Tb and I am getting interface hangs on both Intel and Realtek NICs when the total transferred amount is more than 1TB, sometimes earlier, sometimes later.

 

[atlantis852]

Are you positive it started on 10.9.4? I seem to have had issues earlier as well.

For the port hang issue on my machine, it most definitely started after upgrading to 10.9.4.

Until we we know the actual root cause, it's possible that there could be more than one trigger or issue causing the hangs that so many are having.

My Ethernet port is still stable under v2.4.14 and I frequently hammer the port with large high bandwidth file transfers and streaming. I am not running iSCSI though.

 

[atlantis852]

Are you positive it started on 10.9.4? I seem to have had issues earlier as well.

Besides: I am a bit heavier loaded than you. I am transferring lots of files of 1gb to 3gb size with total amounts from 400Gb to 2-3Tb and I am getting interface hangs on both Intel and Realtek NICs when the total transferred amount is more than 1TB, sometimes earlier, sometimes later.

I only have one Ethernet port on my Maximus VI Extreme. Perhaps something is worse when two ports are present during large file transfers in 10.9.4

Again, until we know the root cause it's hard to say exactly what or why.

 

[atlantis852]

So for clarity, what I am saying is:

1. For my workflow, the Ethernet port hangs started after upgrading to 10.9.4

2. After extensive troubleshooting, the only driver version that made my port stable for my workflow was v2.4.14

3. I am not saying hnak's drivers are the root cause. The root cause is not yet known.

4. I am not saying that rolling back the driver will resolve the issue for everyone. There is clearly more going on than meets the eye. Fortunately, rolling back the driver resolved the hangs for my workflow at least for now.

Hope that helps