spdMerlin spdMerlin - v4.x - speedtest and bandwidth monitoring

[ExtremeFiretop]

Bingo I just recreated the problem but no idea how exactly lol:

I just reconfigured the client to be restricted and then used the external binary instead of built-in

 

[Martinski]

Bingo I just recreated the problem but no idea how:
....
I just reconfigured the client to be restricted and then used the external binary instead of built-in

Something is preventing the router from connecting to the Ookla server via the WireGuard interface.

 

[ExtremeFiretop]

Something is preventing the router from connecting to the Ookla server via the WireGuard interface.

Agreed.

@aex.perez and @fsb

I configured the Wireguard client to have NAT disabled. Inbound Firewall to block. And then added an optional DNS server. With VPN killswitch enabled.

Then configured the spdMerlin script to use the external binary, basically the worst case scenario, and got the error.

So next question would be do you have any of those configured by chance?

 

[ExtremeFiretop]

Slowly been doing process of elimination; Seems to be related to the NAT setting (at least for me)
Whenever I disable it I get the error; right now at this moment my internet works; every other setting is removed, and I'm still getting the error.

As soon as I re-enable the NAT it starts working again:

Here are some "related" sys logs showing the difference between the working and non-working tests a 1 minute apart:

Mar 12 07:32:57 spdMerlin: Speedtest results - Download: 1055.01 Mbps (data used: 1.8 GB ) - Upload: 561.53 Mbps (data used: 540.4 MB )
Mar 12 07:32:57 kernel: br53: port 12(wds3.1.1.52) entered blocking state
Mar 12 07:32:57 kernel: br53: port 12(wds3.1.1.52) entered disabled state
Mar 12 07:32:57 kernel: device wds3.1.1.52 entered promiscuous mode
Mar 12 07:32:57 kernel: br53: port 12(wds3.1.1.52) entered blocking state
Mar 12 07:32:57 kernel: br53: port 12(wds3.1.1.52) entered forwarding state
Mar 12 07:32:57 spdMerlin: Connection quality - Latency: 11.59 ms (0.29 ms jitter) - Packet Loss: 3.0%
Mar 12 07:32:58 spdMerlin: Retrieving data for WebUI charts...
Mar 12 07:32:58 kernel: Assigning idx[100] for wds3.1.1.53
Mar 12 07:32:58 kernel: netdev path : wds3.1.1.53 -> wds3.1.1
Mar 12 07:32:58 kernel: BCMVLAN : wds3.1.1 mode was set to RG
Mar 12 07:32:58 kernel: br58: port 13(wds3.1.1.53) entered blocking state
Mar 12 07:32:58 kernel: br58: port 13(wds3.1.1.53) entered disabled state
Mar 12 07:32:58 kernel: device wds3.1.1.53 entered promiscuous mode
Mar 12 07:32:58 kernel: br58: port 13(wds3.1.1.53) entered blocking state
Mar 12 07:32:58 kernel: br58: port 13(wds3.1.1.53) entered forwarding state
Mar 12 07:33:01 kernel: wfd_unregisterdevice Successfully unregistered ifidx 2 wfd_idx 1
Mar 12 07:33:01 kernel: br0: port 15(wds1.1.1.0) entered disabled state
Mar 12 07:33:01 kernel: br53: port 10(wds1.1.1.52) entered disabled state
Mar 12 07:33:01 kernel: device wds1.1.1 left promiscuous mode
Mar 12 07:33:01 kernel: br58: port 12(wds1.1.1.53) entered disabled state
Mar 12 07:33:01 kernel: Freeing devid[2137] for wds1.1.1
Mar 12 07:33:01 kernel: Freeing devid[2138] for wds1.1.1.0
Mar 12 07:33:01 kernel: br0: port 15(wds1.1.1.0) entered disabled state
Mar 12 07:33:01 kernel: Freeing devid[2139] for wds1.1.1.52
Mar 12 07:33:01 kernel: br53: port 10(wds1.1.1.52) entered disabled state
Mar 12 07:33:01 kernel: Freeing devid[2140] for wds1.1.1.53
Mar 12 07:33:01 kernel: br58: port 12(wds1.1.1.53) entered disabled state
Mar 12 07:33:01 kernel: ^[[0;33;41m[ERROR vlan] vlanIoctl ,668: Failed to delete VLAN device wds1.1.1.52^[[0m
Mar 12 07:33:01 kernel: ^[[0;33;41m[ERROR vlan] vlanIoctl ,668: Failed to delete VLAN device wds1.1.1.0^[[0m
Mar 12 07:33:02 kernel: ^[[0;33;41m[ERROR vlan] vlanIoctl ,668: Failed to delete VLAN device wds1.1.1.53^[[0m
Mar 12 07:33:02 kernel: ^[[0;33;41m[ERROR vlan] vlanIoctl ,668: Failed to delete VLAN device wds1.1.1.0^[[0m
Mar 12 07:33:06 kernel: SBF: dhd1: INIT [00:41:0e:23:06:27] ID 65535 BFW 65535 THRSH 2048
Mar 12 07:33:19 spdMerlin: Starting speedtest using auto-selected server for WGVPN1 interface
Mar 12 07:33:37 SIG_UPDATE: current sig : 2448
Mar 12 07:33:37 SIG_UPDATE: latest sig : 2448
Mar 12 07:33:43 spdMerlin: Error running speedtest for WGVPN1

Edit:

Issuing a service restart_dnsmasq or a service restart_wgc does not resolve the issue as long as NAT is disabled for me.
Switching to the internal (built-in) binary has not resolved it either.

 

[aex.perez]

N/A to me
Not the client, am the server or servers in my case (Opensense, Wireguard and IPsec)

 

[ExtremeFiretop]

N/A to me
Not the client, am the server or servers in my case (Opensense, Wireguard and IPsec)

Okay, no worries.

It would be unrelated to my new Wireguard code then so all I can hope is maybe @fsb can confirm over his Wireguard client if he has those options enabled or disabled while testing.

 

[CaptainSTX]

For others to be safe; I would add a "forceupdate"

/jffs/scripts/spdmerlin develop
/jffs/scripts/spdmerlin forceupdate

Just simply because there can be multiple "4.4.6" versions in the wild depending when they did the develop command against the repo

I had no problems updating using just the first command shown above. Seems to be working fine testing both a single OpenVPN client and two WireGuard clients. I am somewhat skeptical of the results reported for the WG clients however as WG client 1 reports approximately the same download speed as OpenVPN 1 client and the servers for both are in a city 200 miles distant. What is even harder to believe is that WG Client 5 which is in a city 4,500 miles distant is that the download speed reported is 2 -4 times faster than the speeds to the nearby server even with a latency of 155 ms to the distant server while the latency to the nearby WG server is 16 ms.

 

[ExtremeFiretop]

I had no problems updating using just the first command shown above. Seems to be working fine testing both a single OpenVPN client and two WireGuard clients.

Happy to hear! I'm still trying to identify what specific configurations causes that error, but it's functional for me as well, the errors generated don't come from spdMerlin but for the WireGuard client themselves so I believe it is likely a configuration issue for those users.

I am somewhat skeptical of the results reported for the WG clients however as WG client 1 reports approximately the same download speed as OpenVPN 1 client and the servers for both are in a city 200 miles distant.

Actually in my case OpenVPN is just slower, Wireguard is almost always faster for me.

What is even harder to believe is that WG Client 5 which is in a city 4,500 miles distant is that the download speed reported is 2 -4 times faster than the speeds to the nearby server even with a latency of 155 ms to the distant server while the latency to the nearby WG server is 16 ms.

I would just make sure the stats you see are the stats regenerated by WireGuard testing, otherwise it sounds like it's more or less functional if it's giving you such stats from each Wireguard client

 

[Jaime Alvarez]

Maybe it has nothing to do with this issue, but I logged into the router GUI to check spdMerlin and it was not running. I ran a WAN speedtest through the GUI, using a random server but it didn't work. I then SSH'ed in and again I couldn't get a simple WAN speed test to work. In both cases it was claiming something about sockets... I then realized that I had configured a preferred server which wasn't wotking anymore, and that spdMerlin was set to use a preferred server.

Problem solved? Yes. But just to check, I performed another WAN speedtest through the GUI, using a random server, and this time it worked. I don't know why having a non-working preferred server made a difference for the random server test...

 

[Ripshod]

I'm not too sure if this is relevent but I'm speedtesting my NordVPN (wireguard) to Londons servers, and randomly I'm getting errors where the speedtest seems to "freewheel" through a test and returns no data):
Fail:

Mar 12 20:04:15 ripshod spdMerlin: Starting speedtest using Box Broadband Ltd (London, United Kingdom) for WGVPN1 interface
Mar 12 20:04:23 ripshod spdMerlin: SQLite3 failure[spd21]: Parse error near line 2: near ",": syntax error
  pload],[ServerID],[ServerName]) values(1741809857,,,,,,'https://www.speedtest.
                                      error here ---^
Mar 12 20:04:24 ripshod spdMerlin: SQLite3 failure[spd21]: Parse error near line 2: near ",": syntax error
  pload],[ServerID],[ServerName]) values(1741809857,,,,,,'https://www.speedtest.
                                      error here ---^
Mar 12 20:04:25 ripshod spdMerlin: SQLite3 failure[spd21]: Parse error near line 2: near ",": syntax error
  pload],[ServerID],[ServerName]) values(1741809857,,,,,,'https://www.speedtest.
                                      error here ---^
Mar 12 20:04:26 ripshod spdMerlin: SQLite3 failure[spd21]: Parse error near line 2: near ",": syntax error
  pload],[ServerID],[ServerName]) values(1741809857,,,,,,'https://www.speedtest.
                                      error here ---^
Mar 12 20:04:27 ripshod spdMerlin: SQLite3 failure[spd21]: Parse error near line 2: near ",": syntax error
  pload],[ServerID],[ServerName]) values(1741809857,,,,,,'https://www.speedtest.
                                      error here ---^
Mar 12 20:04:27 ripshod spdMerlin: SQLite process reported error(s).
Mar 12 20:04:27 ripshod spdMerlin: Speedtest results -  -
Mar 12 20:04:27 ripshod spdMerlin: Connection quality -  -
Mar 12 20:04:27 ripshod spdMerlin: Retrieving data for WebUI charts...

Success:

Mar 12 20:09:38 ripshod custom_script: Running /jffs/scripts/service-event (args: start spdmerlinspdtest_auto_WGVPN1_)
Mar 12 20:09:43 ripshod spdMerlin: Starting speedtest using auto-selected server for WGVPN1 interface
Mar 12 20:10:07 ripshod spdMerlin: Speedtest results - Download: 501.18 Mbps (data used: 579.9 MB, additional server: 1) - Upload: 62.74 Mbps (data used: 76.0 MB)
Mar 12 20:10:07 ripshod spdMerlin: Connection quality - Idle Latency: 70.05 ms (jitter: 1.01ms, low: 67.43ms, high: 70.92ms) - Packet Loss: 0.0%
Mar 12 20:10:07 ripshod spdMerlin: Retrieving data for WebUI charts...

 

[CaptainSTX]

Happy to hear! I'm still trying to identify what specific configurations causes that error, but it's functional for me as well, the errors generated don't come from spdMerlin but for the WireGuard client themselves so I believe it is likely a configuration issue for those users.



Actually in my case OpenVPN is just slower, Wireguard is almost always faster for me.



I would just make sure the stats you see are the stats regenerated by WireGuard testing, otherwise it sounds like it's more or less functional if it's giving you such stats from each Wireguard client

Just to confirm my suspicions I ran a speed test to the same server using the VPN provider's app running on my PC to the same distant server and the results were roughly half of what the spdMerlin app is showing. Both the router and the app were using WireGuard. I will do some more trials and see if I can determine what is going on.