Hi there,

 

I'm running 4.7 with the hardware in my sig. Lately I'm having many issues with my network loosing connectivity from my main Win 7 machine.

 

I can see this in the log:

 

Dec  1 12:15:17 Tower kernel: r8169: eth0: link down (Network)

Dec  1 12:15:17 Tower ifplugd(eth0)[1378]: Link beat lost. (Network)

Dec  1 12:15:19 Tower kernel: r8169: eth0: link up (Network)

Dec  1 12:15:20 Tower ifplugd(eth0)[1378]: Link beat detected. (Network)

 

I haven't change any hardware, configuration or plugins. The only difference is that I'm building a second server (ESXi 5.1 with unraid 5rc8a, Win 7 and Ubuntu server) so I added to my set up an old gigabit switch (Belkin F5D5141-5).

 

Any ideas? How can I track this down? It's becoming really annoying, I can watch movies almost without any issue from the Media Center in a different room, but each time I directly access the shares from my Win 7 desktop I lost connection like every 15min or so 

Well.. you are experiencing network issues and you have added a network component recently... First thing I would do is take out the new switch to see if the problem goes away.

Ok, I should've explain better 

 

I added the switch like 4 months ago, when preparing the location for the new server, and I'm having these issue in the last 2/3 weeks, and it's becoming more and more frequent.

 

More info, my setup is something like this:

Router -> Switch -> unraid server, esxi server, desktop

 

and when I lose connection to unraid I can still surf the internet, so I though the issue wouldn't be in the switch.

 

However, it doesn't hurt to try with my previous set up, right? 

It could still be the switch - or its power supply - or cables / connectors. 

 

Try swapping ports / cables.  You could also try borrowing a different switch / power supply.  But it could also be the network connection on the server.  The key is to change only one thing at a time and see if there's any noticeable change in the problem.

Jup... that's the route I would go to... Basic troubleshooting, try to change stuff, or take stuff out of the mix, one by one and see what happens, that way you can zoom in on the culprit. I used to have a switch that needed the power to pulled every couple of months, then it would work fine again for a few months.

First thing I tried was to change the cable, it wasn't it  .

 

Then I used my previous switch, and connected directly to it the unraid server and the desktop. I then chained it to the gigabit switch to have internet on my desktop, so now it looks something like this:

Router -> Gigabit Switch -> esxi server, Second-switch -> unraid server, desktop

 

and it's been working with some heavy use (updating XBMC SQL library, accessing shares, Media Companion, etc) for about 40min which is quite a record lately.

 

I'll monitor it closely and let you how it goes.

 

However, this brings me a new question, actually two:

1. How is possible that the switch puts down the connection of a working port? Is there some king of bottleneck to the max traffic, connections, etc, it can handle?

2. And this brings the next question, I was thinking of buying a rack mount switch this xmas, is there any specific specs I should look for (or avoid)?

 

I think it is more likely that it is broken.. You could switch around some ports to see if its localized to one..

no luck 

 

I keep on having connection issues. Any ideas?

Boot without ESXi to test.

I switched off ESXi, and even unplugged the ethernet connections (including IPMI), but I'm still having connectivity issues.

 

The only thing I can think off is the switch, so I bit the bullet and anticipated my x-mas shopping. I saw somewhere in the forums a recommendation for the TP Link SG1016, so I bought the 24 ports version TL-SG1024.

 

I'll report if issues go away.

 

 

each time I directly access the shares from my Win 7 desktop I lost connection like every 15min or so 

One possibility...

 

Make sure the two unRAID servers are using different MAC addresses.  It is frequently found hard-coded and set in the

config/network.cfg file.  That is not an issue until you reuse the motherboard it was on for another device on the LAN where it will override the actual MAC address and result in a conflict on the LAN.

 

If there is a line setting the HWADDR, remove the line on the offending server.

I'm not sure if having two identical MAC addresses on the same lan will cause "linkup/down" issues, but it will cause connectivity issues for sure.

 

Joe L.

Thanks Joe, it wasn't that though, no hard coded MAC addresses.

 

Some more findings. I have a desktop (Win7 32bits) and an ESXi server that has 3 VMs:

#1 - Win7 64bits

#2 - Linux server 12.04

#3 - Unraid 5rc8a

 

After testing with my ESXi server down (here) I switched it back for further testing. My VM #2 (linux) is mounting Unraid 4.7 shares during start up through fstab. Strange thing is that I cannot access unraid 4.7 from my desktop or VM #1 (both are win7 machines) but I can access them from my Linux server.

 

So issue seems to be related to windows , however whenever I cannot access unraid 4.7, I still can access unraid 5

 

I got tired of this, connection kept getting worse and the new switch didn't help at all.

 

So I bit the bullet and moved to 5rc8a a little earlier than I liked to, I drop all together 4.7 and make my ESXi server the production one.

 

So far everything is good, no network issues.

 

Thanks all