v1.4.0rc1-16 Bug reports and comments

[wisp-wand]

Just update my 2 WS-12-250DC and left one on the bench testing. I never saw the spurious voltage and temp noise that we were seeing in the field, when on the bench however I just checked into the switch and there was an exclamation icon and the following:-


Jan 1 00:00:05 netonix: 1.4.0rc10 on WS-12-250-DC
Jan 1 00:00:09 system: Setting MAC address from flash configuration: EC:13:B2:81:xx:xx
Jan 1 00:00:11 system: starting ntpclient
Jan 1 00:00:12 admin: adding lan (eth0.770) to firewall zone lan
Jan 1 00:00:21 dropbear[777]: Running in background
Jan 1 00:00:26 system: starting ntpclient
Jan 1 00:00:26 switch[802]: Detected cold boot
Jan 1 00:00:27 switch[802]: PoE enabled on port 2, PoE Smart is starting cable check
Jan 1 00:00:28 switch[802]: PoE enabled on port 3, PoE Smart is starting cable check
Jan 1 00:00:29 kernel: link state changed to 'up' on port 1
Jan 1 00:00:32 switch[802]: Port 2 cable check results: Open, Open, Open, Open
Jan 1 00:00:32 switch[802]: Port 2 passed PoE Smart cable check, turning on power
Jan 1 00:00:32 switch[802]: Port 3 cable check results: Open, Open, Open, Open
Jan 1 00:00:32 switch[802]: Port 3 passed PoE Smart cable check, turning on power
Jan 1 00:46:28 switch[794]: 48V is out of range (32.6V)
Jan 1 00:46:32 system: starting ntpclient
Jan 1 00:46:40 switch[794]: 24V is out of range (22.6V)
Not sure if the switch took any notice, but certainly looks like we still have spurious stuff coming in, and now on the bench.

W

[Dave]

what load do you have plugged into the switch ports that you turned on, and what is your power supply source?

these are real out of range failures, and with the code change probably not bogus, so we need to figure out whats up so
describing your setup will help.

[sirhc]

If your power source on your bench powering the switch does not have enough amperage this could happen.

What we can look into is having the switch shut OFF in the event the power source does not have enough amperage for the load.

The switch should start shedding or turning off POE ports in the event the powersource is of insufficient amperage to carry the load. We are looking into that.

How many amps does your power supply feeding the switch have, please provide specs on your power source.

But providing us more information on what your power source is, what is its specs and what were you powering on port 2 and 3 would help.
It appears the input amperage became insufficient?

[Dave]

what i meant was it might be possible that your power source in the lab has to low a current limit set on it, or possible bad connections, and if our dc-dc power supply can't get enough current it will still try & output it power to the main board, but it will not be able to keep the 48 volts at 48 volts because it is current starved so it will simply start dropping its output voltage, and that will also effect the 24 volt supply on the switch board.

that is why I am askign about your setup

[IntL-Daniel]

I have higher CPU temp than you...



...but there are still the same conditions..nothing changed. Maybe different RPM vs Temp threshold values?

...just want to inform that since upgrade to 1.4.0rc10 these fan peaks disappeared (back to the same state like with 1.3.9). Also we saw total power consuption peaks over 200W while normally we never saw it before (with 1.3.9) and now with 1.4.0rc10. Maybe the power consuption peaks was the root cause of fan peaks... I did not do double check by downgrading back to 1.4.0.rc9 (our switch is in production).

Edit: ..and YES we have the switch in the plastic box but with the special wholes for fan output and wholes for air input and the box is in the air conditioned container (small room on the roof). The point of my posts were just description of the status of fan/power behaviour in 1.3.9 vs 1.4.0rc9 vs 1.4.0rc10

[IntL-Daniel]

I have still the board at 53C with fan stable at 6500rpm now and hold (1.4.0rc10).

[lligetfa]

No fan on the MINI so cannot comment on that.

Also we saw total power consuption peaks over 200W while normally we never saw it before (with 1.3.9)...

Right after flashing RC10, when I logged in I saw the total peak a couple of times and then settle down. The duration of the peaks were so short that when I clicked on each port to try to determine what was drawing it, the 3 sec graph had passed by. I don't recall seeing that before and have not seen it since.

[wisp-wand]

what i meant was it might be possible that your power source in the lab has to low a current limit set on it, or possible bad connections, and if our dc-dc power supply can't get enough current it will still try & output it power to the main board, but it will not be able to keep the 48 volts at 48 volts because it is current starved so it will simply start dropping its output voltage, and that will also effect the 24 volt supply on the switch board.

that is why I am askign about your setup

Hi Dave, its a 24v SMPSU 2.5AMP. The switch has no load on it at all, just my ether cable plugged in so I can monitor.

It is configured for one of our tower sites, so a couple of the ports have POE enabled, but nothing connected. Screenshot of status page attached

Wayne

[Dave]

wayne

thanks, i am investigating this..for sure the latest rc code has vast improvements, but it seems i have to track down some floating ghost of a bug....

dave

[sirhc]

Also be aware efficiency is so low because there is no load.

It is sort of counter intuitive that the efficiency goes up with load but it does.