v1.3.9 FINAL bug reports and comments

[QuadeHale]

Reyes - did you change something via console/ssh prior to attempting to change the management password via GUI?

I ran in to this a few times; there doesn't seem to be as much 'input verification' in the console as there does in the GUI. I've made changes in the CLI that the GUI doesn't like at all and subsequently the GUI will block me from saving changes until I figure out what's wrong. The worst part is that sometimes the change was made to a part that's not 'exposed' to the GUI yet so you can't even see the red box to see what the offender is.

The 'Invalid Field' dialog/modal should really tell you which field is invalid (specifically) as a starting point, there are a lot of tabs to look through and it might not even be visible, but eventually the CLI should probably get some sanitization/verification.

[sirhc]

Reyes - did you change something via console/ssh prior to attempting to change the management password via GUI?

I ran in to this a few times; there doesn't seem to be as much 'input verification' in the console as there does in the GUI. I've made changes in the CLI that the GUI doesn't like at all and subsequently the GUI will block me from saving changes until I figure out what's wrong. The worst part is that sometimes the change was made to a part that's not 'exposed' to the GUI yet so you can't even see the red box to see what the offender is.

The 'Invalid Field' dialog/modal should really tell you which field is invalid (specifically) as a starting point, there are a lot of tabs to look through and it might not even be visible, but eventually the CLI should probably get some sanitization/verification.

Good idea!

I asked Eric to look at doing something about that

[zhall]

I ran into a 1.3.9 netonix 400B that lost its ability to watchdog reboot/notice anything.

I read someone else posting about how they had to restart the service, so I did that on the command line and it came back to working.



Any news on this watchdog problem?




previous poster on this issue:

I am still having watchdog fail after about 3 days. I have some problems with some sectors that keep losing connection on their Ethernet side. after the 2 day and 22 hour mark of uptime on the switch the watchdog stops functioning.

To get watchdog back I have to ether start the service back up with the CLI or reboot the switch.

Does the watchdog service need a watch dog on it?

[sirhc]

I ran into a 1.3.9 netonix 400B that lost its ability to watchdog reboot/notice anything.

I read someone else posting about how they had to restart the service, so I did that on the command line and it came back to working.



Any news on this watchdog problem?




previous poster on this issue:

I am still having watchdog fail after about 3 days. I have some problems with some sectors that keep losing connection on their Ethernet side. after the 2 day and 22 hour mark of uptime on the switch the watchdog stops functioning.

To get watchdog back I have to ether start the service back up with the CLI or reboot the switch.

Does the watchdog service need a watch dog on it?

Yes it is fixed in v1.4.0rcXX, have you been reading the release notes on the current firmware in development?

I have been sending out PM's telling people they should look at v1.4.0rcXX as it is the best firmware yet.

1) If you run into an issue first upgrade to the latest FINAL version if not using the latest.

2) If problem still exists try the latest RC version

3) If the problem still persists then report the issue in the RC thread comments section as the FINAL version is done with development and we would not go back and make changes to it.

[zhall]

Ah, my mistake, I assumed a final firmware version implied some level of confidence that important features like watchdog aren't broken.

[sirhc]

Ah, my mistake, I assumed a final firmware version implied some level of confidence that important features like watchdog aren't broken.

Well if that was the case then there never would be a need for fixes and patches as you would assume all features would be 100% bug free?

But I am pretty sure all manufacturers find bugs and issue fixes to features on a regular basis?

Recent Example:
UBNT Security flaw that allowed people to gain access to their devices (second time this happened)
UBNT Flow Control Pause Frame Storm Bug
Cisco released several patches in the past year for bugs found in our routers OS
HP releases "bug" fixes all the time

Bugs happen, all you can do is try to fix them as fast as possible.

As far as the ping watchdog issue it took weeks and sometimes months to stop working so if we tested all firmware for a period of time to include this it would take months to test which is great except when people find a bug they do not want to wait months for a fix and we try and release patches ASAP.

[zhall]

That is all true, well thank you for your efforts.

I'll check out the RC release as soon as it seems to be reaching a final stable form.

[sirhc]

That is all true, well thank you for your efforts.

I'll check out the RC release as soon as it seems to be reaching a final stable form.

It is almost done and ready to go final.

The only thing we have had to fix in the next version is the Discovery routine to detect airCAMs and uniFi devices.

As I said and showed in a post in the v1.4.0rc16-rc25 thread I am running v1.4.0rc25 on all my WISP towers and I love the new version but you can wait if you prefer or you can try it on one tower and then another. People testing the RC version is how bugs are found and get fixed but as you can see in the release notes there are a lot of enhancements and fixes since v1.3.9

Plus v1.4.0rc25 has a safety fix for a major Pause Frame Storm bug in the airMAC AC and airFIBER firmware

http://community.ubnt.com/t5/airFiber-B ... true#M3756
http://community.ubnt.com/t5/airMAX-AC- ... true#M8233