network issue

[fred9999]

Hi all,

I have a strange problem. It used to be fairly intermittent, but now happens more often.
I have two desktops (PC1 and PC2), along with other laptops and devices on my network. I have an ASUS router, AC1900 and SMART/RG modem. Running WIN7.

I frequently copy files from PC1 to PC2. Usually works fine, but sometimes it will start the copy, do a little bit, and then halt with the message:
"Problem accessing PC2/folder. Make sure you are connected to the network"
Yet I can see all the devices in network neighborhood, so obviously I'm connected to the network.

When this happens, I notice the same problem occurs when I try to copy from PC2 to PC1.
I can fix the problem by rebooting PC2, but that's not working reliably any more.

Another quirk is, when this error happens, it will also abort any download I happen to be doing at the time, so somehow the error is affecting some other network communication.

Does anyone have any ideas as to where I can start looking - router, switch, cables?
Thanks for any suggestions.

 

[RMerlin]

Monitor the stability of the connection by running ping for an extended period of time (i.e. multiple minutes), and look for lost packets or major spikes in delays.

 

[dosborne]

Does anyone have any ideas as to where I can start looking - router, switch, cables?

Wired or wireless? Managed or unmanaged switch?
Start with the basics, see if the (assuming wired) port shows the correct speed. Try swapping out the cables. If wired, is there a patch panel or wall plate that could also be affecting the quality?

Run a diagnostic tool or long speed test between the 2 systems.

 

[fred9999]

Monitor the stability of the connection by running ping for an extended period of time (i.e. multiple minutes), and look for lost packets or major spikes in delays.

Thanks RMerlin,
I'm kind of a novice... I know how to ping, but how do I know what are lost packets or spikes in delay?

And since the error is active now, since the only way to clear it out is to reboot, it would be good to do the pings now?

 

[fred9999]

Wired or wireless? Managed or unmanaged switch?
Start with the basics, see if the (assuming wired) port shows the correct speed. Try swapping out the cables. If wired, is there a patch panel or wall plate that could also be affecting the quality?

Run a diagnostic tool or long speed test between the 2 systems.

Thanks dosborne,
The connection is wired, and switch is unmanaged - just a simple switch.
Sorry, novice here.
How do I know if the port shows the correct speed? Sometimes when I watch the copy, it seems to start out fast and then slows down.
What kind of diagnostic or speed test?

 

[degrub]

If the two PCs are connected to the same switch directly, the router is not involved. If they are connected to two different switches and the switches are connected each to a LAN port on the router, then both switches and the router switch are in the path. Only for internet access will the router portion be active. The switch in the router is usually independent of the rest of what is going on in the router. So to try to track down if there is a physical problem (port, cable ) one would try to make the problem follow a specific change - like moving a cable from one port to another or replacing a cable with a new one.

Each port has two lights. One that turns on when there is power and the other can be different colors depending on link speed ( 100 Mb/s, 1,000 Mb/s) , usually green or amber depending on speed. Consult the switch user manual for what color means what speed. You can determine which light is the speed light by disconnecting a cable and seeing which light stays on. That is the power light.

You can also go to Control Panel , Network and Internet, Network Connections and right click on the network adapter (Local Area Connection) select STATUS from the pop up menu. This will list the SPEED.

When troubleshooting, i find it easier to watch the port light on the switch for the PC in question to see if it changes color.

Sometimes ports go bad/flaky or the cable goes flaky.

If your switches are all supposed to be Gigabit/s switches look for a port that is not the same color as the rest. That port and the device connected to it will be running slower, most likely.

 

[RMerlin]

Thanks RMerlin,
I'm kind of a novice... I know how to ping, but how do I know what are lost packets or spikes in delay?

And since the error is active now, since the only way to clear it out is to reboot, it would be good to do the pings now?

ping <adresse IP of the other PC> -n 100

At the end of the 100 packets, it will tell you how many packets were lost.

 

[fred9999]

Thanks guys, Rmerlin and degrub...
I'll try your suggestions.

 

[fred9999]

ping <adresse IP of the other PC> -n 100

At the end of the 100 packets, it will tell you how many packets were lost.

here's the results of 2 ping tests

1.first time ping 192.168.1.112 -n 100

Ping statistics for 192.168.1.112:
Packets: Sent = 100, Received = 100, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
Minimum = 0ms, Maximum = 0ms, Average = 0ms

2.second time ping 192.168.1.112 -n 200
do the copy

Reply from 192.168.1.112: bytes=32 time<1ms TTL=128
Reply from 192.168.1.112: bytes=32 time<1ms TTL=128
General failure. *at this point the copy failed
General failure.
General failure.
Request timed out.
Request timed out.
Request timed out.
General failure.
Request timed out.
Reply from 192.168.1.112: bytes=32 time<1ms TTL=128
Reply from 192.168.1.112: bytes=32 time<1ms TTL=128
Reply from 192.168.1.112: bytes=32 time<1ms TTL=128
Reply from 192.168.1.112: bytes=32 time<1ms TTL=128

Ping statistics for 192.168.1.112:
Packets: Sent = 200, Received = 192, Lost = 8
Approximate round trip times in milli-seconds:
Minimum = 0ms, Maximum = 0ms, Average = 0ms

So the failure interfered with all communication

I'm now tracing where the cables are going in order to swap - but I don't really see why it's a cable problem. Why does the problem disappear for a while after a reboot?

Thanks for any observations.

 

[fred9999]

If the two PCs are connected to the same switch directly, the router is not involved. If they are connected to two different switches and the switches are connected each to a LAN port on the router, then both switches and the router switch are in the path. Only for internet access will the router portion be active. The switch in the router is usually independent of the rest of what is going on in the router. So to try to track down if there is a physical problem (port, cable ) one would try to make the problem follow a specific change - like moving a cable from one port to another or replacing a cable with a new one.

Each port has two lights. One that turns on when there is power and the other can be different colors depending on link speed ( 100 Mb/s, 1,000 Mb/s) , usually green or amber depending on speed. Consult the switch user manual for what color means what speed. You can determine which light is the speed light by disconnecting a cable and seeing which light stays on. That is the power light.

You can also go to Control Panel , Network and Internet, Network Connections and right click on the network adapter (Local Area Connection) select STATUS from the pop up menu. This will list the SPEED.

When troubleshooting, i find it easier to watch the port light on the switch for the PC in question to see if it changes color.

Sometimes ports go bad/flaky or the cable goes flaky.

If your switches are all supposed to be Gigabit/s switches look for a port that is not the same color as the rest. That port and the device connected to it will be running slower, most likely.

I checked the cables and they both go to the same switch.
I did several copies (PC1 to PC2), and to my surprise they all ended normally - usually after getting the error, I needed to reboot PC2. Something must have corrected itself.
I watched the switch (Trendnet Gigabit) while the copies were going. The two ports for PC1 and PC2 were flickering quickly, and the other ports flickered slowly - all green. Power light was on.
Twice during one of the copies all 7 lights turned off for about half a second, then all came on again. I thought the copy would fail at this point, but it ended ok. This must be an indicator of something?
I ran the same copy that made the lights turn off and on again, but this time it went normally.

Thanks for any suggestions that come to mind.

 

[RMerlin]

General failure.

This indicates that the network interface itself is getting disconnected - it's more than simple packet losses.

The next step is to look at the computer's network interface, the cable (try replacing it) and the network switch to which it's connected (it might be failing).

wice during one of the copies all 7 lights turned off for about half a second, then all came on again. I

Most likely candidate is the switch itself then.

 

[degrub]

Or the power supply to the switch.
Just buy a new switch. At worst, you will have a spare/growth flexibility.

Try one of the Netgear Gbit unmanaged SMB type switches. Metal case. Almost bulletproof.

 

[fred9999]

Thanks guys. I'll get a new switch and see how it goes.
My new hobby, watching the lights when I copy something.

I have a second issue, and I was just wondering if it might have the same cause.
Sometimes I run downloads at night, and for the last 7 days, they all abort. Sometimes 3 different downloads from 3 different servers, all abort at the exact same minute. The funny thing is, all 7 failures are between 1.21am and 2.57am .
I've been trying the same downloads going to a wireless laptop, and they finish with no issue, while the wired pc fails.

Could this be due to the same switch? If so, why only and always between 1.21 and 2.57.

 

[fred9999]

Or the power supply to the switch.
Just buy a new switch. At worst, you will have a spare/growth flexibility.

Try one of the Netgear Gbit unmanaged SMB type switches. Metal case. Almost bulletproof.

So I was watching the lights while doing some copies, and sure enough, the first time all the lights went out, that copy failed. Later I retried and the same copy finished successfully. So my initial assumption that I needed to do a reboot to clear the error was false. I guess it's just random.
I was lucky enough that BestBuy was on the last day of a sale on Netgear Gbit unmanaged switch, so I bought one.

Never thought a simple switch would fail - I was thinking, no moving parts, why would it fail?
Live and learn.
Will take me a few days to replace the switch.
Thanks again everyone. This has been bothering me for a couple years, finally looks to be solved.

 

[RMerlin]

Never thought a simple switch would fail - I was thinking, no moving parts, why would it fail?

Considering the behaviour, my first guess would be a power supply issue - once the switch starts having a higher workload, it was restarting. The ping test I suggested allowed to confirm that the problem was the link itself that was dropping, not just mere packet losses.

Thankfully, Gigabit switches are fairly cheap these days, now that Multigig switches have started to become affordable.

 

[fred9999]

Well, I replaced the switch, and things seemed to be fine for a few days, and I stopped watching the lights when copying files.. The lights show 1G connection.
Yesterday, it happened again. Started a copy from 1 pc to the other pc, and it failed almost immediately.
I tried the same copy a few minutes later with no problem.

What would be the next likely culprit - the cables between the pc and the switch?

Thanks.

 

[RMerlin]

Well, I replaced the switch, and things seemed to be fine for a few days, and I stopped watching the lights when copying files.. The lights show 1G connection.
Yesterday, it happened again. Started a copy from 1 pc to the other pc, and it failed almost immediately.
I tried the same copy a few minutes later with no problem.

What would be the next likely culprit - the cables between the pc and the switch?

Thanks.

Cables would be the next thing to check. There is also the possibility of a bad NIC.

Weirdest issue I've experienced in my work was a customer's network going down every few minutes, lots of packet losses as well. By process of elimination I tracked down the issue to a bad NIC in an old computer. Unplugging that computer stabilized the whole network. So, sometimes the cause might be in a totally separate device connected to the same network.

 

[Ripshod]

Talking of bad NICs, there can be a big difference in drivers. If you're using the NIC/motherboard manufacturer's drivers you could try the generic drivers from the chip maker, and visa versa. Try to avoid the default windows drivers. Instabilities can show anywhere.