R700 100% cpu /usr/bin/matrix after upgrade from 9.12.1.0 build 140 to 9.12.2.0 build 204

  • 1
  • Question
  • Updated 3 years ago
After upgrade from 9.12.1.0 build 140 to 9.12.2.0 build 204, 10 of our 39 R700 have 100% CPU utilization for at lease 20 minutes after the upgrade.
If we reboot the R700 the CPU is still 100% CPU (0 clients has been connected).

In the support file we noticed that the /usr/bin/matrix process was responsible for that.  

We opened a case for that 00414781 a P2.

After one hour after the upgrade we decided to downgrade to version ZD3000 9.12.2.0.101, this version has the same behaviour.

Does someone have this same behaviour with this software version ?

Do you have details about the processes /usr/bin/matrix 
Photo of Arthur Hulsman

Arthur Hulsman

  • 19 Posts
  • 0 Reply Likes

Posted 3 years ago

  • 1
Photo of Dave Watkins

Dave Watkins

  • 65 Posts
  • 13 Reply Likes
Yup, I've had a case open for this for R600's for a while now. I've been told a fix has already been developed and will be available shortly, but I'm trying to get it backported to a custom firmware I'm already running to fix another issue. The timeframe for this fix to be released I was told was 9.13 - sometime in June.

Initially (I've had this going on for a while now) I didn't have any client connectivity issues related to this, but they have started cropping up now and then now
(Edited)
Photo of Bart Oudhuis

Bart Oudhuis

  • 4 Posts
  • 0 Reply Likes
Hi Dave, what kind of controllers are you using? We see this problem with the ZD3000 only, our ZD1200 doesn't seem to have the high cpu load on APs (small test environment with 2 R700).

Yesterday I received a call from ruckus support, they informed me that the problem should only occur if you use the ZD3000. A fix should be released somewhere next month.
Photo of Dave Watkins

Dave Watkins

  • 65 Posts
  • 13 Reply Likes
We're on redundant ZD3000's so that doesn't surprise me. I was just about to see if I could hurry our ticket up as I'm now starting to get reports of users not being able to connect, which at this point i'm attributing to this. We've got just less than 200 R600's all affected so thigns are a little painful at the moment
Photo of Arthur Hulsman

Arthur Hulsman

  • 19 Posts
  • 0 Reply Likes
Herewith the Engineering Reference Number:ER-3716,SCG-48499
Photo of Michael Brado

Michael Brado, Official Rep

  • 2826 Posts
  • 397 Reply Likes
Fix has been checked into DevEng code, and will be incorporated in next 9.12.2 refresh (target May 6), 
and SmartZone 3.4 releases.
Photo of Bart Oudhuis

Bart Oudhuis

  • 4 Posts
  • 0 Reply Likes
Hi Michael, can you give us some insight on the fix release? Would be great to know if we can expect a release within a week.
Photo of Dave Watkins

Dave Watkins

  • 65 Posts
  • 13 Reply Likes
If you have a ticket open you should be able to get it now. I've had it running here for a few days without issue so it looks to be resolved