v1.3.0 & v1.3.1 & v1.3.2 FINAL bug reports and comments

[sirhc]

After upgrading 16 of my switches to 1.3.0, all was well until 12 hours ago when I lost a WS-24-400B which was supplying 1 mimosa, 1 24GHz HD airfibre, 1 5GHz airfibre and 3 powerbeams, the 48v line reverted to default off in the gui but the 24v stayed on, I will replace with 1.3.1 and see how it goes

I am not sure I follow what happened?

1 port turned itself off or multiple ports turned themselves off?
Are you saying that ports 1-3 or 4 powering 48VH shut off?

Go to the Device/Status Tab and please tell me the Board Rev

Also please post a screen grab of the Status Tab

[designweb]

To clarify everything was normal on all config inc port settings, however the 3 x 48v radios listed were powered down, the poe settings which were on 48v changed themselves to 'off' as if the power had never been configured. I have now reverted the two airfibres to their own POE blocks and re-enabled the Mimosa from the switch. I do have another identical switch which had been upgraded to 1.3.0 carrying out the same task, 2 airfibre and one mimosa which carried on unchanged

Total Throughput
Power Usage: 50.8 W
95.64
TX Mbps
14.87
TX Kpps
95.20
RX Mbps
14.73


Current Time 19/08/2015, 21:27:19
Uptime 13:06:11
CPU Usage 28%
Memory Usage 44.28 MB
MAC Address ec:13:b3:11:0b:e2
IP Address xxx.xxx.xxx.xxx
Firmware Version 1.3.0
Board Rev C
Power
Board 48V 50.5 Vdc
Board 24V (Ports 1 - 12) 24.7 Vdc
Board 24V (Ports 13 - 24) 24.7 Vdc
Board 3V 3.3 Vdc
Thermal
Board Temp 36 C / 97 F
CPU Temp 59 C / 138 F
PHY Temp (Ports 1-12) 49 C / 120 F
PHY Temp (Ports 13-24) 35 C / 95 F
Cooling
Fan 1 Speed 2310 RPM
Fan 2 Speed 2400 RPM
Fan 3 Speed 2400 RPM

[sirhc]

Did you check the log to see if someone turned them off or if anything else was registered in the log?

I have this version on 13 towers now with no issues

No I had wanted to so a screen Grab of the Main Status Tab with the graphic of the switch and the watts being drawn on each port but since you pulled the units from the switch that is impossible now.

[sirhc]

We are looking at a routine in v1.3.X that was put in mainly for the WS-12-DC unit that if it detects the input voltage dips below 43V for 5 seconds it will turn off 48V POE ports to protect against damaging 48V POE devices because we simply pass through the switch board input voltage for 48V POE out.

I believe the routine puts a log entry (I am checking with Eric but he is off line) in the switch log so please check if you did not reboot or clear the log.

Remember our AC power supplies are 50V +/- and our DC Power supplies convert 9-60V to 50V to feed the switch board.

The switch board then down converts 50V to supply 24V devices.

The WS-12-DC model is meant for people who already have 48V-53V conditioned sites but the unit will run on 37V and still supply 24V POE devices.

This "MIGHT" be it, would be nice if you still have the log.

It could have been a fluke, we are checking into it as well.

[sirhc]

I just noticed that you were running v1.3.0, PLEASE upgrade to v1.3.1 as there were changes in v1.3.1 that are important.

Also check the LOG if you can and report back.

[designweb]

The earlier report suggested that 1.3.0 was problematic , it was the first upgrade that caused one of my switches to fail and because it failed and took many customers offline I upgraded to 1.3.1 which unfortunately wiped the log but got the customers back on.

Lesson learned from your comments on the routine for the DC unit is we cannot use the switch POE to power the incoming, link else as happened yesterday we cannot reach the switch at the tower to recover and fix any problems like buggy firmware or be able to re enable POE on other ports which defeats object running everything from POE switch in remote tower or where there is DC renewable energy only.
Would it be possible to disable that routine on the port in use for incoming feed perhaps? Ellis my techie (another member) and me are the only people with switch access and neither of us changed anything on the config

[sirhc]

We do not know this is what happened?

I was just trying to "guess" as to a possible cause, as I said I have many towers running this and I have not had a single issue.

Since that routine checking input voltages is "new" to v1.3.X I was thinking of that as a "possible" cause, it may not be.

So far no one else has reported any issues like this.

Next time if you ever have an issue be sure to look at the log file for a cause. Also the log file will tell you if someone at your office turned them off by mistake which is also a possibility as it logs this activity to a source IP.

I think you are jumping the gun calling this a "buggy" firmware. You had a single event, you do not know the cause, we are speculating at this point.

[designweb]

Apologies I was merely reporting my experience that it was a one off amongst all of my twenty switches I am not a developer I am a user and my use of buggy firmware term was following your thoughts of a possible cause not meaning the firmware generally has a problem, also amongst my switches only one out of the two that have three 48v radios powered, shut its 48v power line down not the 24v side, I will check the logs in future and send those as well should there be anything to report.
In this case as I mentioned there are only two of us that access these switches and neither of us changed any config, I have full confidence in the switches and will continue to purchase and use them to replace many more in my network.

[petelarsen]

After upgrading one of my switches (WS-12-250A) to 1.3.1 I can no longer power any 48V devices. The ports that were set to 48V are now set to off and the 48 V poe option is grayed out on each port.

[Jknight]

Petelarsen, the user above me and I are talking about the same switch
Here are the logs:

Aug 20 07:00:14 dropbear[1348]: Running in background
Aug 20 07:00:27 switch[1379]: Disabling PoE power on port 5 due to Input Voltage too low for 48V operation (12.9 Vdc)
Aug 20 07:00:29 switch[1379]: Disabling PoE power on port 6 due to Input Voltage too low for 48V operation (12.9 Vdc)
Aug 21 15:32:19 UI: Configuration changed by 209.33.215.142
Aug 21 15:32:19 UI: Port 5 Enable: changed from 'Enabled' to 'Disabled'
Aug 21 15:32:19 UI: Port 6 Enable: changed from 'Enabled' to 'Disabled'
Aug 21 15:32:59 UI: Configuration changed by 209.33.215.142
Aug 21 15:32:59 UI: Port 5 Enable: changed from 'Disabled' to 'Enabled'
Aug 21 15:32:59 UI: Port 6 Enable: changed from 'Disabled' to 'Enabled'
Aug 21 15:33:09 UI: Configuration changed by 209.33.215.142
Aug 21 15:33:09 UI: Port 5 Link: changed from '100M-F' to 'Auto'
Aug 21 15:33:09 UI: Port 6 Link: changed from '100M-F' to 'Auto'
Aug 21 15:33:23 UI: Configuration changed by 209.33.215.142
Aug 21 15:33:23 UI: Port 5 Link: changed from 'Auto' to '100M-F'
Aug 21 15:33:23 UI: Port 6 Link: changed from 'Auto' to '100M-F'

The 7AM referenced in the log from 8/20 was the time I upgraded the switch. Ports 5 and 6 are connected to M5 Ti's, which cannot get 48v.

Image below is what the drop menu looks like, with the greyed out 48v