v1.3.8 FINAL bug reports and comments

[adairw]

Maybe on next release we will increase the default to 120 seconds?

No, Chris, The correct answer is, we will increase the time to 120 seconds that only works if you hit the test button.

[TheHox]

Setting up a new mini at a small tower, changed the name, IP, login and password. The whole tower went offline.

It finally came back a few minutes later, it reverted all changes back to the original settings. Problem is it changed the login name, but not the password.

The login was reverted back to "admin", but the password I just put in was saved. Took for a bit to figure that out..

[adairw]

Chris, are you listening? The auto revert thing is annoying. LOL

Test button or bust.

[sirhc]

Chris, are you listening? The auto revert thing is annoying. LOL

Test button or bust.

Are you reading the v1.3.9rcX Release notes?????


viewtopic.php?f=17&t=1232

We addressed that for you almost immediately!


v1.3.9rc8 - Released 12/15/2015
- Fixed UBNT Discovery name
- Ping Watchdog now disabled if port is disabled
- Changing Safe Mode Timer and ability to disable Safe Mode
- Ability to reset counters per port instead of all
- Log file formatting
- UI Enhancement for MAC table
- Add clear stats command to CLI

[adairw]

I had not. That's still not what I'm asking for though.

[TheHox]

Yea there are just so many firmwares out there (netonix/ubnt/airmaxM/airmaxAC/M&C-mixedmode/DFS/5.1/AF5X/) I just stay behind a version or two.

I'll be sure to check out 1.3.9 final though :)

I'm not sure what the rc has relating to the test/apply mode. But, The way UBNT has it works perfect for me, if that is the plan for 1.3.9 I'm all in.

The way 1.3.8 has it, whatever it did dropped the whole tower and all customers because I changed the name of the switch, the IP address and the password.

[lligetfa]

I'm not sure what the rc has relating to the test/apply mode. But, The way UBNT has it works perfect for me, if that is the plan for 1.3.9 I'm all in.

The way 1.3.8 has it, whatever it did dropped the whole tower and all customers because I changed the name of the switch, the IP address and the password.

Chris/Eric is reluctant to go with the test button for whatever reason.

What there is in the latest RC, is the option under Device > Configuration to set a higher value or to disable it completely.

Selection_005.jpg (39.22 KiB) Viewed 6666 times

[sirhc]

I fail to see the value in the "Test" Button over the way we do it? Please explain the benefit?

You can set the Revert timer to whatever period of time in seconds within reason you want or "0" for Disable.

When you hit Save/Apply if the UI/CLI loses communications with the switch and does not regain it within the defined period of time it reverts to the previous config, if it connects back to the switch in the allotted time it assumes all is good.

[bradsdb]

I just got my first Netonix switch in the mail. I have a WS-12-250-DC. I think this is running 1.3.8.

I started to program the switch, eg I set the new password, ip address, and ntp server settings. Then after applying the settings a pop-up came up that would not go away saying something about no more changes because config was out of sync and to click here to reload configs. I clicked the link and it seemed that the switch was reloading configs but the pink pop-up message kept appearing at the top of the page. So, I did a reboot from the web page menu. After the reboot the pop-up did not come back.

Having to reboot the switch to sync the configs is a bug. You should not have to reboot equipment like ubiquiti gear has to be rebooted :)

After about 5 minutes - 20 minutes of the switch sitting on the bench (with no changes made) the switch would not respond to ping. I turned on wireshark, and began captureing packets. I plugged another device into the switch and was not able to ping the other device.

I connected to the switch over serial port, and did a show status... below is the result. The CPU was never lower then 18%, sometimes 26%+.

Netonix Switch# sh status
System
------
Current Time: 12/31/1969 18:30:46
Uptime: 5445.910156 seconds
CPU Usage: 26%
Memory Usage: 42.63 MB
MAC Address: ec:13:b2:81:26:02
IP Address: 100.124.17.73
IPv6 Address: fe80::ee13:b2ff:fe81:2602
Firmware Version: 1.3.8
Model: WS-12-250-DC
Board Rev: E

Power
-----
Board 24V: 24.9 V
Board 3V: 3.3 V
Power Supply Input Voltage: 24.0 V
Power Supply Input Current: 0.3 A
Power Supply Output Voltage: 49.5 V
Power Supply Output Current: 0.2 A

Thermal
-------
Board Temp: 38 C
CPU Temp: 55 C
PHY Temp: 59 C
Power Supply Board Temp: 34 C
Power Supply Heatsink #1 Temp: 23 C
Power Supply Heatsink #2 Temp: 23 C
Power Supply Controller Temp: 37 C

Cooling
-------
Fan 1 Speed: 2400 RPM


I then did a "show log" and noticed some ntpclient messages. I have seen in some linksys manages switches that they crash 1 month later due to a misconfigured sntp value. Maybe ntp not being reachable is causing the switch to crash?

The settings I have put into the switch wont work until I connect the switch to a network fyi.

Here is the log output.

Dec 31 17:08:17 system: starting ntpclient
Dec 31 17:10:17 system: starting ntpclient
Dec 31 17:12:18 system: starting ntpclient
Dec 31 17:14:18 system: starting ntpclient
Dec 31 17:15:58 STP: set port 1 to discarding
Dec 31 17:15:58 kernel: link state changed to 'down' on port 1
Dec 31 17:16:17 system: starting ntpclient
Dec 31 17:18:07 STP: set port 1 to discarding
Dec 31 17:20:06 kernel: link state changed to 'up' on port 1
Dec 31 17:22:17 system: starting ntpclient
Dec 31 17:24:06 STP: set port 1 to discarding
Dec 31 17:26:06 kernel: link state changed to 'down' on port 1
Dec 31 17:28:17 system: starting ntpclient
Dec 31 17:30:08 STP: set port 1 to learning
Dec 31 17:32:06 kernel: link state changed to 'up' on port 1
Dec 31 17:34:17 system: starting ntpclient
Dec 31 17:36:06 STP: set port 1 to forwarding
Dec 31 17:40:17 system: starting ntpclient
Dec 31 17:44:18 system: starting ntpclient
Dec 31 17:46:17 system: starting ntpclient
Dec 31 17:48:18 system: starting ntpclient
Dec 31 17:50:18 system: starting ntpclient
Dec 31 17:50:27 STP: set port 1 to discarding
Dec 31 17:50:27 kernel: link state changed to 'down' on port 1
Dec 31 17:52:17 system: starting ntpclient
Dec 31 17:54:06 STP: set port 1 to discarding
Dec 31 17:56:06 kernel: link state changed to 'up' on port 1
Dec 31 17:58:17 system: starting ntpclient
Dec 31 18:00:08 STP: set port 1 to learning
Dec 31 18:02:06 kernel: link state changed to 'down' on port 1
Dec 31 18:04:18 system: starting ntpclient
Dec 31 18:06:06 STP: set port 1 to forwarding
Dec 31 18:06:42 init: starting pid 282, tty '/dev/ttyS0': '/bin/ash --login'
Dec 31 18:08:06 kernel: link state changed to 'up' on port 1
Dec 31 18:10:18 system: starting ntpclient
Dec 31 18:12:06 STP: set port 1 to discarding


Is there a manual or anyplace I should start reading more about this switch?

For a brief period of time the switch responded to ping after several minutes of not responding to ping. I have several known good pieces of equipment I can plug into the switch to be 100 sure its not my laptop.

I am available to trouble shoot this more, is there a tech support number?

[sirhc]

Hello

There are over 9000+ switches in service, you would think that if this was a problem with the switch you would not be hearing good things but all bad things.

Now with that being said I suggest you install v1.3.9rc14 as it has several fixes. I would read the release notes for v1.3.9rcX and see what bugs were fixed.

Now the log entries about the NTP service starting over and over again is because you either have an incorrect NTP server configured or it is unable to get to the NTP server and will continue to continue to try and reach it and until it does you will see this entry every so many seconds. What would you prefer it do, check once and if it fails for whatever reason stop trying?

Now the CPU utilization means nothing to do with the switching functions of a switch, all the CPU does is run the UI/CLI and stats daemons as the packet switching is handled in the switch core by dedicated processors just like any other switch (not including software switches where all packets are handled by the CPU).

As far as you losing pings to the switch I would look into your network settings on your computer or look for an IP conflict on your network (meaning something else is at 192.168.1.20 already) which would also make sense about the drop down complaining about config.

The UI is pretty straight forward but we have a video tour of the switch here https://www.youtube.com/watch?v=8JvBEAD4MFM

In the video I go over all the settings and even give you a tour of switches in service at my WISP