v1.4.0 Bug Reports and Comments

[slowe]

I am having issues where SFP modules will just stop working. Everything is running fine then we just loose activity. Link stays up but no traffic is passed. In the logs I am seeing 'SFP: FAULT on SFP 4, resetting...' anybody else see this? Reboot usually clears everything up

Thanks!
Spencer

[Banana Jack]

We are fighting with this since upgrading from 1.3.9 to 1.4.0

May 25 17:11:32 STP: set port 3 to discarding
May 25 17:11:34 STP: set port 3 to learning
May 25 17:11:34 STP: set port 3 to forwarding
...etc.
Br.
Carsten

Our switches were doing the same thing on 1.4.0 with STP enabled Carsten.

Kind regards, Glenn

[sirhc]

Yea there are 2 bugs in v1.4.0

I am pushing up v1.4.1rc2 that should fix this and then if it does we will push 1.4.1 FINAL as these are 2 bad errors.

Please test these and let me know if it fixes your problem so I can push v1.4.1 FINAL

[sirhc]

@Banna Jack

PLEASE try v1.4.1rc2 and let us know if this fixes your problems as soon as possible.

[Hightech]

Thanks for the fast update, I have updated all my switches to v1.4.1rc2, i will report back tonight, so fare it semse to fix the STP issue we have seen in our network.

Br.
Carsten

[Banana Jack]

@Banana Jack

PLEASE try v1.4.1rc2 and let us know if this fixes your problems as soon as possible.

Very impressive fast fixes, thanks sirhc. I downloaded and installed 1.4.1rc2 on my bench WS-6-MINI with Loop Protection disabled and STP enabled. I connected it to our production network and everything seems to be working fine, so that's great.

When 1.4.1 final version is released I will install it on our production switches.

Thanks
Glenn

[sirhc]

When 1.4.1 final version is released I will install it on our production switches.

Thanks for getting back to us but this is why we are fixing v1.4.0 just days after release because I can not get people to install the RC versions and test them.

Put it on a couple easy to get to tower switches and make sure!

If people find a bug report it and if we can confirm it then work with us to allow Eric access if needed to solve the problem.

People need to get more involved with the RC versions to help debug them.

I put them on my WISP network but there are so many different configurations out there and no-one is willing to help debug them not even our "Associates"

[Banana Jack]

Thanks for getting back to us but this is why we are fixing v1.4.0 just days after release because I can not get people to install the RC versions and test them.

Put it on a couple easy to get to tower switches and make sure!

If people find a bug report it and if we can confirm it then work with us to allow Eric access if needed to solve the problem.

People need to get more involved with the RC versions to help debug them.

I put them on my WISP network but there are so many different configurations out there and no-one is willing to help debug them not even our "Associates"

Thanks, I appreciate where you're coming from. I can't speak for the WISP community in general, but from my own point of view I haven't previously upgraded our Netonix switches with Beta/RC releases for several reasons, including:

1. I had presumed that the releases were primarily intended for testing and I hadn't signed up for any beta programme with Netonix, nor had Netonix vetted me as a tester who they would actually trust (!). Plus the releases seem to be very frequent, which is great in terms of the effort that your developers are clearly putting in, but quite time consuming in making and monitoring regular deployments from a tester/user point of view.

2. The Netonix firmware upgrade process is one of the slower ones that I have seen on network appliances generally (not a criticism, just an observation) and so upgrading all our switches is a reasonably time-consuming process. More importantly it involves quite a lot of network downtime which is not what our customers want especially when the aim of the exercise is not to fix any specific problem or usually to give any very noticeable improvement from our customers' point of view.

3. Previous Netonix firmware releases have been very stable in my experience, and so I do think that in some respects, "if it ain't broke, don't fix it". On the other hand, I do like to support the policy of continuous improvement, and I am grateful for such an approach.

I hope you don't mind my candid reply, but I thought it would be best to give you honest answers. Having said all that, I think it's really commendable that you guys are so dedicated to creating an absolutely top-notch product even though some people might think, "it's only a switch". Also, now that I know about the MANAGER program which I understand can schedule upgrades, I imagine that my point 2 no longer applies. As a result, I take your comment on board and will certainly try to test future releases and report back any problems if it helps with the development process.

Kind regards
Glenn

[sirhc]

This was not directed at just you Glenn but the WISP Switch community as a whole.

Everyone only wants to run FINAL versions but then there is a greater chance that we miss something as there are so many network configurations out there and my WISP is like a cookie cutter which means all my towers are setup the exact same so I will not catch everything.

Actually our upgrade process is faster then most other switches also our switches stay running during the upgrade process and only stop switching packets during the reboot process which is less than 30 seconds and we leave the POE ports powered up during the reboot process so wireless links do not power cycle during the reboot process.

We do have Associates but most (not all) seem to only want run final versions and rarely participate in the RC shake downs.

Since we pretty much build specialized equipment used just by WISPs or commercial businesses there are not really any civilian users here to speak of like over on the UBNT forums so we assume the majority of our users are at least capable of participating in the RC process when they have spare time so we allow everyone access to the RC versions in hopes that a few each time will have the ability to test and comment on an RC version every now and then. Since we have almost 1500 users you would think that an average of 5-10 could spare some time to help the community on each release?

We do not expect WISPs to sit down and debug all the firmware features but if enough people just simply install it on an easy to access switch and make sure it works with their network topology then if enough WISPs do this as all typologies are different then we would get a large sample of tested typologies.

We have so many things we want to do such as QOS and other advanced features but we need to get the current feature set stable before we move forward.

[Banana Jack]

@sirhc

That was a great post. You should make it a sticky or put the thrust of its sentiments somewhere prominent on the site. Then newbies like me will get a hold of the situation more easily.

In terms of the firmware update process itself, it's great to learn that the downtime is minimised and that (as you mentioned earlier in the thread) a disconnected browser session during the update doesn't matter. These two points should be strongly highlighted, as I had never known about them previously, and a firmware update process on a production network can be a potentially stressful task at the best of times.

Thanks
Glenn