Bad NIC driver? Or bad NIC? or neither

[Ayeohdee]

FreeNAS 9.1.0
Motherboard : EVGA nForce 750i SLI FTW
Ram : 8gb OCZ DDR2 800
CPU: Q9505
NIC: Intel PWLA839GTL 10/100/1000 pro 1000 GT
Storage: 3x 2TB WD RED + 1 120GB Muskin SSD

It seems any time I transfer a large data file (500+mb) my nas will seem to lock or not respond for some time via the GUI and shell. On an older version on freenas I was getting Interrupt storm on irq21 which was my network device. Now that I have upgraded to 9.1.0 here is what I am seeing now during this sudden connection drop / massive lag spike.

 

[dlavigne]

That error deals with disks, not the NIC. What does zpool status look like?

 

[Ayeohdee]

That's sad to hear =( and odd. I do not doubt you at all it's just the strange thing about that is if i use my onboard realtek lan I dont have that issue. Assuming I can get an IP to hold with the realtek I can transfer as much as I want and all goes well, its just as soon as I use my Intel card I have this issue. The zpool shows an error, I cant tell you word for word right now but something along the lines of a device from the pool is missing and or cannot be opened. I think that is because I was using an SSD for cache on the zpool in 8.3 but for reasons had to remove it when importing this pool into 9.1.0. I had healthy status before upgrading to 9.1.0. I certainly could be wrong, but this has been happening since I got the Intel NIC. As I said before if I use the Onboard realtek NIC I do not have this issue. That's whats confusing to me about it being an error with this zpool.

 

[SweetAndLow]

Make sure all sata cables are connected and power is connected. Something might have wiggled lose.

 

[Ayeohdee]

I'll check on the connectors for sure. I am actually looking to get a PCIE SATA controller since a couple of the sata ports on the mother board are cracked but I did the best job I could snugging them fit when I built the system. All of my data on the server as far as I can tell is still accessible and not corrupted but here is the zfs warning I am getting . The volume Storage (ZFS) state is ONLINE: One or more devices could not be opened. Sufficient replicas exist for the pool to continue functioning in a degraded state. I believe I am getting that error because the volume had an SSD for caching that I removed before I imported the volume it into 9.1.0, I could be wrong though.

 

[m0nkey_]

To me, it looks like the storage controller is about to take a ****. The device at ada2 just got booted too. My guess your pool is now in a degraded state.

If that was my box, I'll be backing up everything on it and looking for new hardware.

 

[cyberjock]

Honestly, I'd seriously consider shutting down your system until you can get your PCIe controller. Things are pretty precarious based on the messages you shared.

 

[Ayeohdee]

I'll certainly take all of your advice. I don't have any precious data on the server that I am really worried about but I will continue searching for a known compatible PCIe controller and start from scratch. Thank you all for the fast replies.

 

[Ayeohdee]

So I cannot say I rightfully "solved" my problem but based on the fact that these error's were occurring only when using my intel NIC I decided to run down to a local shop and purchased a different NIC card (tp-link tg-3486) and also payed a super super premium price for some obviously used sata cables for good measure. I replaced the NIC, installed some new cables and I backed my data up, wiped the drives, created a new volume and I am currently in the middle of a 200gb transfer to the server and everything is going smooth.

 

[Ericloewe]

To me, it looks like the storage controller is about to take a ****.

That's just another day for nForce users.