v1.3.6 FINAL bug reports and comments

[sirhc]

Please post comments and reports bugs with v1.3.6 FINAL in this thread

As always be careful to roll out new versions, upgrade switches that are easily accessible until you verify that the new firmware works with your network design and desired feature set.

We try to test as much as possible but it is impossible to test every possible network configuration.

FIXES
- Changed accepted voltage ranges on some models

ENHANCEMENTS
- More verbose I2C error handling and reporting

KNOWN ISSUES
- Ubiquiti Discovery Protocol is still flaky
- IPERF is still a work in progress
- WS-6-MINI and WS-8-250-AC and WS-8-250-DC always see a reboot as cold which you may have your first POE Port Shut OFF after reboot.
We are working on this ISSUE

Released 10/28/2015

FIXES
- Fixed flow control on dual SFP ports
- Fixed PoE power being turned off when port is disabled
- Fixed problem with MAC table not being displayed
- Fixed selecting 24V when bulk editing PoE settings
- Fixed bug parsing times in timed shutdown periods for PoE on DC models
- Fixed login problem when changing credentials name without also changing password
- Fixed DHCP fails it falls back to a static address but continues to try to get a DHCP address
- Fixed setting SNMP community string in CLI
- Fixed problem with changes to Switch Name not taking effect immediately

ENHANCEMENTS
- Added autonegotiation to dual SFP ports
- Added error condition notifications to the UI
- Added meters to UI
- Added sysDescr to SNMP with model number
- Added missing indexes to SNMP (fan, voltage, poe status and temp)
- Added SMTP Local Hostname option
- Added root bridge ID and port role to STP tab
- Added PoE Smart check for suspected 24V devices when selecting 24VH POE
- Added option to use web interface on HTTP as well as HTTPS

KNOWN ISSUES
- Ubiquiti Discovery Protocol is still flaky
- IPERF is still a work in progress
- WS-6-MINI and WS-8-250-AC and WS-8-250-DC always see a reboot as cold which you may have your first POE Port Shut OFF after reboot.
We are working on this ISSUE

Released 10/28/2015

NOTE:
Our switches are designed to continue to run even when they are damaged. Sometimes the switch can take damage to sensors from a surge from the tower or ground potential shift between the tower and the cabinet damaging the current sensor on a port. Since all the current, voltage, and temp sensors share the same I2C bus this can cause other sensors to report erroneous values.

Prior to v1.3.5 we did not do as good of job at alerting the user to out of range readings, therefore after upgrading to v1.3.5 or newer if you are now seeing voltages, watts, or temperatures out of range this may be an indication that your unit took damage in the past and you were unaware of it.

This is why I say to always make sure your Tower Ground is Bonded to your Electrical Service ground and to run a ground wire up the tower and to each radios on the tower. I have a good post HERE about proper grounding methods.

Another possible cause of "false" out of range readings can be from an incompatible SFP module. Since the SFP cages also share this same I2C bus to relay information about the SFP module to the system some incompatible SFP modules "can" disrupt normal I2C communication. Try removing the SFP module and see if the readings return to acceptable ranges. If you find an SFP module that is causing issue try another brand of SFP module. We are working on a way to eliminate this interference from some SFP modiule manufacturers and may offer a firmware in the future to fix this issue with these modules.

[dan_]

In my test lab on a WS-8-250-DC default config, with 1 poE device connected:
When upgrading to 1.3.6 from 1.3.4 i found that the switch disabled PoE on port4 which had a "tycon tp-din web monitor" attached. This happend on the following reboot as well.
Log entry said something about "link detected on port 4, disabling PoE"

After a couple of reboots it now seems ok:

Code: Select all

 
Jan 1 09:30:25 switch[766]: Detected cold boot
Jan 1 09:30:26 switch[766]: PoE enabled on port 4, PoE Smart is starting cable check
Jan 1 09:30:31 switch[766]: Port 4 cable check results: Open, Open, Short, Short
Jan 1 09:30:31 switch[766]: Port 4 passed PoE Smart cable check, turning on power
Jan 1 09:30:33 STP: set port 4 to discarding


My concern is that if we upgrade our production equipment, we will then lose connection to the switch because it disables PoE on the port connected to our AF backhaul.

Upgraded again and confirmed the issue:

Code: Select all

 
Jan 1 09:30:29 dropbear[1331]: Running in background
Jan 1 09:30:32 system: starting ntpclientJan 1 09:30:34 switch[1358]: Detected cold boot
Jan 1 09:30:34 STP: set port 4 to discarding
Jan 1 09:30:35 switch[1358]: Port 4 has link, disabling PoE


Has anyone else seen this issue?

[sirhc]

I saw this happen once on one of my tower upgrades and 2 other people have reported this over the past year.

It happens because the switch thinks it is a cold boot (power cycle) not a warm boot (software reboot) after a firmware upgrade or software UI reboot request.

You are the 4th person to ever "report" this including me (the posts are on the forum here), it may have happened more often and was simply never reported?

This is why we now report in the log of it thinks it is a cold or warm boot.

We will investigate further, that you for your observation/testing and the post to notify us.

[sirhc]

My concern is that if we upgrade our production equipment, we will then lose connection to the switch because it disables PoE on the port connected to our AF backhaul.

So we just had a conference call discussing your reported issue Dan.

We have determined that the way the switch currently decides if the switch is booting from a warm software reboot such as from a firmware upgrade or a cold boot from a power cycle has a "small" potential of incorrectly determining it is a Cold Boot when indeed it is a warm boot. In such a case the 1st port with POE configured may get turned OFF, but not all the POE ports would be turned off.

The outcome of our conference call just now on this issue is that we have decided to implement a more full proof way of determining the boot state (cold or warm). It is a relatively small change so we will implement this change and test it out then release a new version (v1.3.7) once we have tested it to be a better method.

Once again thank you.

[dan_]

it says "detected cold boot" After the firmware update and after the reboot.


It seems that I can replicate the issue over and over by (re)-upgrading to 1.3.6

Code: Select all

Oct 29 09:35:32 dropbear[1314]: Running in background
Oct 29 09:35:37 switch[1336]: Detected cold boot
Oct 29 09:35:37 STP: set port 4 to discarding
Oct 29 09:35:38 switch[1336]: Port 4 has link, disabling PoE


Turning the PoE back on manually and
Reboot 1 (GUI)... same problem:

Code: Select all

Jan 1 09:30:25 switch[721]: Detected cold boot
Jan 1 09:30:26 switch[721]: Port 4 has link, disabling PoE
Oct 29 09:44:12 UI: Port 4 PoE: changed from 'Off' to '48V'
Oct 29 09:44:15 switch[721]: PoE enabled on port 4, PoE Smart is starting cable check
Oct 29 09:44:20 switch[721]: Port 4 cable check results: Open, Open, Short, Short
Oct 29 09:44:20 switch[721]: Port 4 passed PoE Smart cable check, turning on power


Reboot 2 (GUI) - OK

Code: Select all

Jan 1 09:30:25 switch[728]: Detected cold boot
Jan 1 09:30:26 switch[728]: PoE enabled on port 4, PoE Smart is starting cable check
Jan 1 09:30:31 switch[728]: Port 4 cable check results: Open, Open, Short, Short
Jan 1 09:30:31 switch[728]: Port 4 passed PoE Smart cable check, turning on power


[dan_]

Great to see such quick response thanks.

In such a case the 1st port with POE configured may get turned OFF, but not all the POE ports would be turned off.

Once again thank you.

I guess having PoE on port 1 and our AF24 Backhaul connected to it would be the "1st port with POE configured"... oh oh....

I look forward to an updated firmware and i'll test it on this device where i am currently able to reproduce the symptoms.

Cheers,
Dan

[sirhc]

Dan,

Could you do a factory default then re-set up your LAB.

Do not export and import config, we want a fresh config file.

Let us know if you try this.

[dan_]

No worries.. Will do to make sure

FYI it was factory default except

Poe on 4
nTP server set
Then upgraded from 1.3.4

[rgunderson]

All my WS-8-250-AC units accepted 1.3.6 fine but my WS-12-250B was filling up the log with "voltage should not be zero" on ports 3 & 4 so I downgraded back to 1.3.3. The voltages never did work right on this unit. Should I RMA? See attachments.

WS-12-250B-Image1.gif (189.41 KiB) Viewed 9743 times

WS-12-250B-Image2.gif (147.61 KiB) Viewed 9743 times

[rgunderson]

I take it back, a couple of my WS-8-250-AC units starting acting up after my last post. It may be unrelated to the firmware as one of the boxes is a unit I have previously RMA'd twice. The logs say they are detecting a bridge loop but after spending a long night without sleep I have determined there are none. The suspect switch and the other one (victim?) are strangely taking turns dropping pings while pinging simultaneously. Ive got a screen capture in both video and image showing this and it is easily reproduced. I'd upload it but the forum doesnt like it