network issue

[Tyskern]

Hi all.

 

A network question.

 

Our display computer is connected via a couple of switches, allowing tcp and udp traffic.

And it basically works fine. but if I reboot one of the switches it does not.

 

Actually i dont get an error message, which I would expect..

 

After the display does not update while changing positions in timeline to view different clips.

If i do an update ctrl-d. it will update according to the timeline, but I’m unable to play anything.

If I go offline, then online again, it will also update to the correct pos. in timeline, and I can press play to view a clip, BUT only if i press play within ca one second after going online. Count to four, press play, nothing happens.

 

Now I do a Stage->Remote access on the display computer, do absolutely nothing, close the window, and then go online again.

Then we are back to normal and everything works.

 

I’ve had some similar problems earlier, and learned that WO uses IP traffic for update/online, and timeline changes are done with UDP.

The ports we used then did allow for this, but we changed the switch ports to allow any kind of traffic, and then WO worked properly.

 

So what kind of network does WO actually use. Are there some additional(more than ip or udp) network sendt, and why does a remote access solve it, as one would assume an off/online with thumbs up and green lights would be enough?

 

 

[Miro]

Are you using a managed switch, router or unmanaged/dumb switch? Seems like the UDP packages are lost somewhere on the way. Do you use any firewalls?

[Tyskern]

its a managed switch, but ports used are managed to allow everything. no firewalls..

It actually works 99% of the time.

[jfk]

its a managed switch, but ports used are managed to allow everything. no firewalls..

It actually works 99% of the time.

 

I have encountered managed switches where the exact same claims were made,

and yet UDP blocking was occurring.

This was easily confirmed by substituting an unmanaged switch for a test.

 

When the IT guys were faced with that evidence,

they discovered that even though all ports were open on the managed switch,

UDP flood (aka UDP storm) protection was still active.

(A UDP storm may also be known as a Denial Of Service [DOS] attack.)

Since WATCHOUT broadcasts UDP packets on a continuous cyclic basis (albeit very low bandwidth)

the managed switches were interpreting it as an attack and blocking the UDP transmissions.

ymmv