Unstable Mesh Network

  • 1
  • Question
  • Updated 6 months ago
Hello,

I try to be short. ;)

Problem
9x Unleashed APs (T300 and T301s) on a big parking area.
2x of them are Root-APs. 7x are Mesh-APs. No Wireless Clients. Only IP-Cameras connected per Cable with the Mesh-APs.
Issue: Network unstable.

What i have done with Ruckus Support
Contacted Ruckus Support about 6 Month ago. Support first thought it is because the Mesh-APs are jumping between the Root-APs. Suggestion: Set a fix Mesh-Uplink-Route.
Before doing so i decided to disconnect one Root-AP. So no hopping between Root-APs possible anymore. This did not solve my issue. We discarded this Suggestion.
After that Support checked the Logs and thought is must be a firmware bug. Because the Network was unstable it was hard for us to do an Update. But after restarting the Root-APs the Network was suddenly stable (Ping stable at 1 - 2 ms) for one whole week and we deciced to do the Update.
So we updated from 200.1 to 200.3. But Problem was back again.
After that Support checked the Logs again and told me there are Interference within the 5GHz Band. I should do a Site Survey. Case closed. No Support.

What i have done by myself
After that i tried to investigate this a little bit further by myself.
Because the Installation Site is about 800km away from me and the APs are mounted about 10m high it's not that easy to try something.
Because the Network was stable about one Week after restarting the both Root-APs i did not believe in any Interference within the 5GHz Band.
So i restarted my Root-APs again and since that the Network is stable again.

Where is the Difference?
I am not sure where the Problem is. This is because i need your help and experience.
After restarting the Root-APs the Master Role changed for a short period of time but at the end the Master was the same AP again. So it did not play a role which AP the Master is.
But when the Network was unstable there where 5 Mesh-APs connected with the Master-Root-AP and only 2 Mesh-APs with the Standby-Master.
After restarting it is exactly vice versa.
Maybe it is a load issue (CPU/RAM) or in the end a (till now) unkown Firmware-Issue?

What do you think?
Is there something i can try to find the Root-Cause?
Any Suggestions welcome.

Many Thanks.
Kind Regards
Marco
Photo of Marco Eichstetter

Marco Eichstetter

  • 148 Posts
  • 7 Reply Likes

Posted 6 months ago

  • 1
Photo of Monnat Systems

Monnat Systems, AlphaDog

  • 760 Posts
  • 163 Reply Likes
Marco.
network is un-stable is too generic. what happens when network is unstable??
how unstable network impacts you customer? what stops or does not work?
Photo of Marco Eichstetter

Marco Eichstetter

  • 148 Posts
  • 7 Reply Likes
Sorry.
Unstable means that i have a poor Ping Response Time (>400ms) which also differs much.
Also my Mesh APs are going offline and did not response to a Ping (Timeout). Timeout also differs. Sometimes its only 3 Pings which get lost. Sometimes the Mesh-AP is not reachable for multible Minutes.
Within my Unleashed GUI i am getting a lot of Heartbeat loss.

Thanks.
Photo of Alan Rothenstreich

Alan Rothenstreich

  • 63 Posts
  • 11 Reply Likes
Marco, are the root APs changing channels?  I am not familiar with unleashed but on our mesh networks connected to a Ruckus ZD, we use a command through the CLI to disable channel fly after 60 minutes of uptime.  If the root AP detects interference and changes channels, the mesh APs drop and reconnect.  Also, what kind of switches are you connected to?

Alan
Photo of Marco Eichstetter

Marco Eichstetter

  • 148 Posts
  • 7 Reply Likes
Yes. The APs changing channels from time to time.
Since the Restart of the Root-APs there where several Channel Changes. No Issues till now. So changing the channel seems not to be an issue.

Both Root-APs are connected to a PoE Switch. Don't know the Vendor. I think its a TP-Link Gigabit Switch which is web-managed.
Photo of Marco Eichstetter

Marco Eichstetter

  • 148 Posts
  • 7 Reply Likes
No Ideas for me?
Photo of Alan Rothenstreich

Alan Rothenstreich

  • 63 Posts
  • 11 Reply Likes
Do you limit the APs that can mesh together or are you using Smart Mesh?  An issue can arise when a meshed AP believes it has internet service or there is a change of the subnet.  Is your internet connection a dynamic or static IP?  We have all of our APs connected to a Ruckus Controller so much more visibility into what is occurring that way. 
Photo of Marco Eichstetter

Marco Eichstetter

  • 148 Posts
  • 7 Reply Likes
SmartMesh Settings are at the Default Settings. I think there will be generated a Warning if my Mesh-APs using more than 3 Hops. In my Case all my APs connecting directly to one of my Root-APs. So each Mesh has 1 Hop.

Internet is a dynamic IP.

Any other ideas?
Thanks.
Photo of Alan Rothenstreich

Alan Rothenstreich

  • 63 Posts
  • 11 Reply Likes
Is there any correlation between when the public ip changes and the stability issue?  I only use static ip addresses for remote access and stability.
Photo of Marco Eichstetter

Marco Eichstetter

  • 148 Posts
  • 7 Reply Likes
No. There is no correlation. Sure, for remote-access a static IP is always the best. But in my case there is no remote-accesss. All my APs and Controller are in the same Subnet.