v1.4.9 Bug reports and Comments

[vnce]

i can not downgrade as well.. 1.4.9 will not take anymore then the 300 vlans we have in it, it just wont save,, it wont get the time from our ntp server where 1.4.5 had no problems with it, and now I can not add customers with out the addition of vlans as well. there seem to be alot of bugs in this release. I have not only used different browsers but different computers linux, mac windows, chrome firefox safari, dosent work.. i cant even tag a vlan to a port and save that it just reverts to the last known..

so yes downgrade would be nice as this is in production.

[IntL-Daniel]

There is some issue with QOS in 1.4.9. Since upgrade to 1.4.9 twenty four hours ago (5 switches in our network) many customers started to complain about not possibility to watch online videos...that sound is not synchronized with video. When I disabled all QOS rules in all switches, the problem was 100% solved. Other 100% solution is downgrade to 1.4.8 and keep all QOS rules active as before. So we stay on 1.4.8 ...

Example of our QOS tab is here:

netonix-qos.jpg

I'll be on vacation until the 30th, but I can look into it further when I get back.

Any update on this?

[ZeRoc0ol]

Upgrading main switch at the data center - WS-24-400A
partially failed ... switch did not reboot back up ... we had to drive at 5 AM at the data center to reboot the switch manually - not good.

[sirhc]

Upgrading main switch at the data center - WS-24-400A
partially failed ... switch did not reboot back up ... we had to drive at 5 AM at the data center to reboot the switch manually - not good.

This "can" happen on rare occasions however I have never seen this at my WISP and I upgrade a lot more than most people as I test a lot of code not released.

How long was it up?
What version was you upgrading from?

[jpaine619]

<sigh> 3rd lockup in a month on 1.4.9. And the second time I've had to recover with the paperclip method. I had a switch lock up AGAIN today, drove to the tower, cut the power to the switch, counted to 30, powered the switch back up. ...... NOTHING.

Switch was totally unresponsive. So, I cut the power again, stuck in the paperclip, etc.. etc.. etc.

Then spent the next half hour re-configuring it.

Never had a lockup on 1.4.8, nor any of the rc releases leading up to it.

Something is wrong with 1.4.9

[sirhc]

<sigh> 3rd lockup in a month on 1.4.9. And the second time I've had to recover with the paperclip method. I had a switch lock up AGAIN today, drove to the tower, cut the power to the switch, counted to 30, powered the switch back up. ...... NOTHING.

Switch was totally unresponsive. So, I cut the power again, stuck in the paperclip, etc.. etc.. etc.

Then spent the next half hour re-configuring it.

Never had a lockup on 1.4.8, nor any of the rc releases leading up to it.

Something is wrong with 1.4.9

I am running v1.4.9 on all my switches no lock up. If there was something wrong like this you would think a LOT of people would be reporting this.

Go ahead and downgrade to v1.4.8, my guess is issues will remain.

I would love to investigate this but simply saying I have a "lock up" does not help especially when I do not see this.

My "guess" is something else has changed on your network? Maybe you upgraded to newer airMAX AC radios or airFIBER radios and your seeing a Flow Control Packet lock or something?

Nothing that I can think of changed between v1.4.8 and v1.4.9 to cause this.

NOTE: IGMP snooping over VLANS, MSTP, and ERPS are still being developed so use these features at your own risk.

FIXED/CHANGED
- Fixed ubiquiti Discovery TAB with new ubiquiti firmware; not sure if this fixes where some have reported issues when doing a UBNT Discovery from a UBNT router causing issues with the Netonix switches?

- Fixed changing VLAN List in MSTP configuration
- Fixed MSTP root message
- Fixed to only show OCP controls if board has OCP

ENHANCEMENTS
- Added ifHighSpeed from IF-MIB
- Added option to enable both SNMP v1 and v2c

KNOWN ISSUES
- Some language templates need help - please private message Eric Stern to help
- IGMP snooping over VLANS, MSTP, and ERPS are still being developed

Released 12/27/2017

Maybe you could do some testing to help.
Did you attempt to connect with a console cable before defaulting or rebooting to see whats up?
WHat model switches was this?
Need a lot more info to even begin to help.

CLICK IMAGE BELOW TO VIEW FULL SIZE

[dcox]

I'm not sure if this has been posted yet, but snmpwalk will not complete proper walk of OIDs with v1.4.9, worked just fine in 1.4.8 and previous
example:

v1.4.9 ws-6-mini:
root@voyager:~# snmpwalk -On -v2c -c public 10.76.5.9 .1
.1.3.6.1.2.1.1.1.0 = STRING: Netonix WS-6-MINI
<SNIP>
.1.3.6.1.2.1.17.7.1.4.2.1.2.0.150 = Gauge32: 150
.1.3.6.1.2.1.17.7.1.4.2.1.2.0.149 = Gauge32: 149
Error: OID not increasing: .1.3.6.1.2.1.17.7.1.4.2.1.2.0.150
>= .1.3.6.1.2.1.17.7.1.4.2.1.2.0.149

v1.4.8 ws-6-mini:
root@voyager:~# snmpwalk -On -v2c -c public 10.76.4.9 .1
.1.3.6.1.2.1.1.1.0 = STRING: Netonix WS-6-MINI
<SNIP>
.1.3.6.1.6.3.16.1.5.2.1.5.3.97.108.108.1.1 = INTEGER: permanent(4)
.1.3.6.1.6.3.16.1.5.2.1.6.3.97.108.108.1.1 = INTEGER: active(1)
.1.3.111.2.802.1.1.6.1.7 = No more variables left in this MIB View (It is past the end of the MIB tree)

[dcox]

Update:

Looks like the reason the walk fails is because the OIDs start printing out of order...

[J-K]

Every router I have? All my unix servers? Or more accurately, everything that's not "WISP" gear. :)

Syslog server does no good if the switch isn't passing traffic or if you're troubleshooting anything that involves losing connectivity.

Thank you for making me laugh!

[garrison]

On the Device > Configuration > SMTP section, the Password field cannot handle the # symbol in the password. Using a password with a # in it will result in a segmentation fault error. I was able to reproduce the error with a fake password, even. Entering a password that uses # and pressing the Test button will cause it to happen. This has been an issue for past firmware versions as well, but I did not realize the # was causing it or I would have notified the forum about it earlier.