Why proactive multicast performance test fails when health check passes?

[Bleeder]

vSphere/ESXi 6.0, all patches applied.  Multicast performance test is failing with 0.00 received bandwidth.  All other health checks including the multicast assessment and proactive storage performance tests are passing.

What can I check to determine the problem here?

[elerium]

I have the same issue as well with health check and multicast perf result also tests with 0.00 bandwidth. The VSAN Network status check comes back normal in the General->VSAN page. My switch statistics show everything functional, my switch is a Cisco 6513 switch.

Everything on my VSAN/vSphere 6.0 seems to be in good working order though so I haven't bothered to open a SR about it.

[paudieo]

Hi Guys

Paudie from VMware here

I did a quick check internally and the dev team want your help with this

Can you either file an SR or supply us with support bundles (vCenter and ESXi diagnostic files ) with time and date you executed the tests

Please let us know

Thanks

[zdickinson]

How many nodes do you have?  I had the same issue.  I believe the performance test requires 4 hosts, and I only had 3.  Thank you, Zach.

[Bleeder]

That's unfortunate.. only 3 nodes here.

[paudieo]

Hi Guys
based on your feedback I reduced my  4 node cluster to 3 nodes

this is my result, see screen shot

I am not sure where the three node information came from . I do not see >3 VSAN nodes as a requirement on http://www.vmware.com/files/pdf/products/vsan/VMW-GDL-VSAN-Health-Check.pdf

[paudieo]

As I mentioned earlier, it would be great to get a SR opened with VMware or upload diagnostic bundles up here please

[zdickinson]

I opened up a support case and the tech told me that 4 nodes were required.  Possible I was given bad information.  Thank you, Zach.

[paudieo]

Hi Zack

Can you fire me across the SR # so I can understand what's happening

/Paudie

[zdickinson]

It was SR:  15666846305  It was opened for the MTU check failing.  During the troubleshooting of that it was mentioned that 4 hosts were needed for the performance check.  I think, it's been awhile.  Don't hold me to it   Thank you, Zach.

[paudieo]

Hi Zack

I had a look at the SR, the details are inconclusive  or were not supplied within the case  i.e., > 3 node VSAN was not logged on the ticket.

I can request the logs to be retrieved but not sure if that will gain me anything, as it seemed the case was more  related to vDS health-check

As you said it was a while ago lets just move on..

Hi Bleeder‌‌ ,

Since you have a live one , if ‌you want to pursue this (cause I do) can you file an SR.
If you don't want to file an SR please  make logs available (50MB file attachment limit) here or please make available via some other means

As I said earlier , Dev team assured me that there is not a requirement for > 3 node VSAN node cluster for proactive MTU test to work , and I've verified this in my lab via screenshot already supplied.

/Paudie

[zdickinson]

Thanks for checking, my memory must be hazy.  Have a great day!  Thank you, Zach.

[kellino]

Just found this thread and wanted to add my experience, as I can share that we are experiencing the same error in a cluster with 3 hosts.

[GEJN]

I have a four node cluster. Every Health Check and Proactive Test is passed except for Proactive Test/Multi Cast, which fails for two nodes and the same nodes every time I run the test. vSAN Health plugin is updated to 6.01 released August 3.

At the beginning of the implementation it was only one node, which fails - it was the node, which is passed now. I reinstalled the Health Plugin at all four ESXi hosts and the behaviour changed as shown in the screen dump

running tcpdump shows multicast traffic between nodes.

I have followed instruction from this blog: Virtual SAN Troubleshooting: Multicast | VMware vSphere Blog - VMware Blogs

The ESXi host are configured as follows, where vSAN traffic is running at 10Gbit ports 

• HP DL380p Gen8
• 2*CPU E5-2697v2
• 7* 300 GB 10K SAS Disk - HP Enterprise
• 1* 200 GB SSD Disk - HP 691864-B21
• 384 GB ram 24* 16gb 2rx4 pc3-14900r (DDR3-1866)
• NIC – 1gbe 4-port flr adap
• NIC – 10gb 2-port 560 sfp
• HP 32 GB SDHC Enterprise kit 700136-B21

[jonretting]

What does your multicast/snooping/querier VLAN situation look like on your network?  Thanks, -Jon

[GEJN]

We use snooping querier - option two in this blog from Cormac

http://cormachogan.com/2014/01/21/vsan-part-15-multicast-requirement-for-networking-misconfiguration...

My colleaque in the network department have conducted the setup of the VLAN and I don't have access to the configuration.

[jonretting]

If it's possible I would sync up with him and find out what options can be done to troubleshoot. I can't really say what his options will be, especially since I can only assume you have a significant amount multicast traffic on the network. In my opinion the ball is in your net admin's court. -Jon Retting

[Bleeder]

Ok, I think I need to clarify my question.  I just want to know why the health check passes, even though the multicast performance test fails.  It doesn't seem like the health check should be passing with such a result.

[srodenburg]

I'm in the same Boat. As far as i can tell i did everything right concerning Multicast on the network and the VSAN runs fine and dandy, but the Multicast Performance Test always fails with 0.00 MB/s.

[JohnNicholsonVM]

Can you post the switch port configuration/IGMP snooping config for the VSAN VLAN etc?

Also what is the make/model of your switches?