ER605 mass client disconnection events
I am experiencing clients frequently being dropped from the LAN on ER605. Sometimes this is individual clients. Sometimes mass disconnection events when the entire LAN client group gets disconnected for around 10 minutes, as illustrated in this chart -
ISP is fixed wireless on ER605 WAN port with PPPoE connection and failover to Netgear 4G modem on WAN2 with dynamic IP connection.
Client disconnections are not associated with ISP failures / failover. This has been monitored by ISP for weeks and the ER605 is maintaining a stable upstream connection. The issue is on the ER605 / LAN.
ER605 connects to local network via three TP-Link switches in different parts of the building TL-SG1016PE, TL-SG108E and TL-SG1008P. The 8 port switches are currently on separate LAN ports on the ER605 but previously were connected through the 16 port switch with same issues experienced. I moved them to the ER605 LAN ports as I initially thought it might be an issue with the 16 port switch and wanted to test with it isolated. Disconnections persist across whole of network, not just the clients connected through the 16 port switch.
OC200 hardware controller (firmware 1.24.0) is connected via the 16 port switch.
There are two non-TP Wi-Fi access points connected, but issues are with Ethernet connections.
Network speeds via Ethernet are also very degraded. For example, the ISP connection is 120Mbps but speed test on a device connected to the TL-SG108E via Ethernet is only achieving download speeds of around 30Mbps with all other network switches connected, or around 75Mbps if I disconnect the TL-SG1016PE and TL-SG1008P from the ER605 before running the test. Streaming from YouTube or Netflix / other streaming services experiences frequent stalled stream and disconnections.
Aside from some IP reservations for specific clients and the failover to WAN2, the configuration of the network is as per the ER605 default with a single LAN and default VLAN.
Grateful for any suggested solutions.
- Copy Link
- Subscribe
- Bookmark
- Report Inappropriate Content
I should clarify that it is not actually dropping clients. Rather, it seems to be a LAN failure. This is the current list of client uptimes showing several Ethernet connections that have been up for more than 3 days, despite the previous graph showing zero clients for half an hour today -
- Copy Link
- Report Inappropriate Content
do you have to restart router to get everything online or get it online by itself?
how is the network speed? is it normal or slow ?
- Copy Link
- Report Inappropriate Content
@MR.S it recovered itself. No restart. Speed test tonight after this event was good, but we did have some intermittent stalled streams around 7:30pm.
- Copy Link
- Report Inappropriate Content
According to the graph, it shows that everything is offline during that time.
Can you make sure of several facts?
1. No devices appeared as Disconnected in the Devices tab during that time.
2. During that time, is the Internet still working?
3. Like I said, you need to confirm with the ping with a computer. Or nothing you say convinces me. The load balancing(failover) would not take over the connection unless the online detection still gets replies.
Verify the 8.8.8.8 and accessibility to the gateway IP by ping. I need both results. ping ip -t
4. When this happened, after doing all the tests above, connect the PC directly to the modem and verify ping 8.8.8.8, internet availability and speed.
- Copy Link
- Report Inappropriate Content
@Clive_A sorry for the delay in responding. I have not had a time to shut down other users to run the ping tests until recently.
Ping test results as follows:
--- 192.168.0.1 ping statistics ---
30 packets transmitted, 30 packets received, 0.0% packet loss
round-trip min/avg/max/stddev = 0.769/0.971/1.388/0.100 ms
--- 8.8.8.8 ping statistics ---
30 packets transmitted, 30 packets received, 0.0% packet loss
round-trip min/avg/max/stddev = 53.418/55.793/57.867/0.722 ms
- Copy Link
- Report Inappropriate Content
Thanks for posting in our business forum.
Is this the result you did during the "mass client disconnection" period?
- Copy Link
- Report Inappropriate Content
- Copy Link
- Report Inappropriate Content
I've been examining the ER605's events logs and this may be relevant.
It seems that the switches are being intermittently disconnected from the LAN, as shown in the log extracts pasted below.
The curious thing is that the TL-SG1016PE is affected more, even though it is sitting in the same data cabinet as the ER605 and is connected via a brand new cable. I could understand perhaps the TL-SG108E being more affected as it is in another part of the building and using an older cable through the roof, but not the other way around.
The TL-SG1008P doesn't show up as a client, so I can't attach a screenshot of its log, but when I look at the logs of the clients connected to that switch, I can see that they have disconnected and reconnected at exactly the same times, which indicates that it is the switch that is disconnecting, not the individual clients, and the times align with some (but not all) of the same times at which the TL-SG108E is disconnecting and reconnecting. Those two switches are on separate ports of the ER605 (with separate cables from different parts of the building).
It seems to me that there must be something going on either in the ER605 hardware or network configuration that is causing the switches to be disconnected.
- Copy Link
- Report Inappropriate Content
Thanks for posting in our business forum.
RecapsGallion wrote
I've been examining the ER605's events logs and this may be relevant.
It seems that the switches are being intermittently disconnected from the LAN, as shown in the log extracts pasted below.
The curious thing is that the TL-SG1016PE is affected more, even though it is sitting in the same data cabinet as the ER605 and is connected via a brand new cable. I could understand perhaps the TL-SG108E being more affected as it is in another part of the building and using an older cable through the roof, but not the other way around.
The TL-SG1008P doesn't show up as a client, so I can't attach a screenshot of its log, but when I look at the logs of the clients connected to that switch, I can see that they have disconnected and reconnected at exactly the same times, which indicates that it is the switch that is disconnecting, not the individual clients, and the times align with some (but not all) of the same times at which the TL-SG108E is disconnecting and reconnecting. Those two switches are on separate ports of the ER605 (with separate cables from different parts of the building).
It seems to me that there must be something going on either in the ER605 hardware or network configuration that is causing the switches to be disconnected.
So, unmanaged switches would not show up in the log if they disconnect.
About the 1016PE and 108E pop up in the log as disconnected, I really suggest you examine the Ethernet cable. Swap with a different or a new Ethernet cable (better test it working great before you try it on the switch).
I am worried that this is not a problem with the router. Based on my experience, it could be the switch. For this switch, it is not smart enough to block a loop connection if there is any.
Highly recommend you test with a shorter and working cable, and run several devices for a well and monitor its status. Will that log still pop up?
- Copy Link
- Report Inappropriate Content
Thanks for your reply. The SG1016PE is connected by a brand new 30cm cable, and I have tried changing the cable in previous attempts at diagnostics.
But your reference to the SG1016PE being unable to block loops got me thinking - I did have a Synology NAS connected via both Ethernet ports, without link aggregation configured. That might be causing a loop. I have unplugged one cable and will see if that helps.
- Copy Link
- Report Inappropriate Content
Information
Helpful: 0
Views: 1939
Replies: 21
Voters 0
No one has voted for it yet.