Failing to reconnect to GW1000 over a few days

[broadstairs]

On the latest update I am seeing this error for some days now

Code: Select all

2023-03-25 07:17:46.371 DoCommand(CMD_GW1000_LIVEDATA): Error - Unable to read data from the transport connection: Operation on non-blocking socket would block.

just over night I have had quite a large number of errors. It seems like I first saw this around 23rd March. Looking at the Ecowitt data from this gateway there is a gap from 07:35 to 07:55 this morning but the rest of the time looks OK. My other gateway reporting to Ecowitt and WD ran fine during this time with no data gaps in either. So no issues with the internet or wireless connections.

Any ideas?

Stuart

Live data error logs.zip

[broadstairs]

Just checked the logs again and this error is still happening intermittently, however it seems at present to be recovering better.

Stuart

[HansR]

For the GW1100 problem: did you try to reboot the GW1100?

[broadstairs]

No as when I looked at the GW1000 the lights were normal.

Stuart

[HansR]

No as when I looked at the GW1000 the lights were normal.

What have lights got to do with it if the sockets may fail once in a while?
If you reboot you just lose one cycle and see if the error recurs after that within a day. Just worth a try.

[broadstairs]

Usually in my experience the lights do look different if there is a connection issue. Anyway the last error of this type was at 11:50 this morning and has not happened since. so I'll leave it for now and see what happens.

I'd be interested in knowing exactly what causes that message and what it is attempting to indicate.

Stuart

[mcrossley]

The logs posted do not cover the period 07:35 to 07:55 on the 23rd, but...

The gateway seems to be giving the odd random error, MX immediately reconnects, and everything is OK again for a while.

The other times it seems to not even allow reconnections for a while (indeed it seems to disappear off the network) like this that persisted until 23:49...

Code: Select all

2023-03-24 23:17:18.980 Error opening TCP port: No route to host

Once the connection to the upload server has failed as well...

Code: Select all

2023-03-23 22:29:01.312 PHP[Int]: Error - No route to host
2023-03-23 22:29:01.312 PHP[Int]: Base exception - No route to host
2023-03-23 22:29:01.606 PHP[Int]: Error - No route to host
2023-03-23 22:29:01.606 PHP[Int]: Base exception - No route to host
2023-03-23 22:29:01.606 PHP[Int]: Upload process complete

[broadstairs]

Strange that the GW1100 continues to work OK on the wi-fi and upload to Ecowitt just fine, WD is also uploading to the website during these times as well so must be just the GW1000 failing to connect correctly although most of the time there does not appear to be any gaps in upload to Ecowitt except for the short time around 07:17.

I'll do some more digging into router logs.

Stuart

[mcrossley]

For the most part the gw is refusing to accept connections, it may well still be transmitting up to the Ecowitt cloud.

[mcrossley]

The times you get no route probably means the MAC address has timed out from the lookup tables because it hasn't been seen for a while.

[davepaterson]

I am getting a similar problem. This morning @ 04:35,36 & 37 there was no log entry.
Attached diag.file.

20230413-152000.zip

CMX runs 24/7 on a small Windows 11 laptop with a GW1100C, would it be possible to add a ping command to check Laptop to router 198.168.0.1 and also to the GW1100 on 198.168.0.1 to check network between the 3 components? Which would give a check, are they all connected?
Nothing was shown on the CMX Program screen

[davepaterson]

opps: GW1100 on 198.168.0.1 should be GW1100 on 198.168.0.17

[water01]

Try using https://www.nirsoft.net/utils/multiple_ping_tool.html

[davepaterson]

Looks a good prog. But can it in added to CMX??, The standard CMD Ping will bring up the DOS screen which will show the reply/or no reply to check later, I don't do 24/7 myself and when data is lost I am not going to run a prog. at 4 in the morning.

[water01]

Program once loaded pings each nominated IP address every 30 seconds and maintains a log so you can monitor dropouts etc. Keeps going until you tell it to stop and is better than a DOS prompt as it will ping all 254 IP addresses in your IP range if you want it to.