VMware Cloud Community
PeterRaschLager
Enthusiast
Enthusiast
Jump to solution

Every day at about 4:20, 7:20, 10:20, 13:20, 16:12, 19:20, 22:20 and 1:20 all 8 Hosts go 100% CPU

Hello.

How do I trace which process, scheduled task or what ever is causing that?

It does so even without vcenter appliance running or NSX Manager or Veeam Backup.

vCenter appliance 

  • Version:7.0.2
  • Build:18455184

Esxi Hosts:

 

  • Hypervisor:VMware ESXi, 7.0.2, 18426014
  • Model:ProLiant DL360 Gen10
  • Processor Type:Intel(R) Xeon(R) Gold 6134 CPU @ 3.20GHz
  • Logical Processors:32
  • NICs:6
  • Virtual Machines:53

There are no scheduled task visible on the Hosts and two scheduled tasks on the vcenter that runs once per day (VMware vSphere Lifecycle Mananger)

We use Nimble HF40 as datastore but there are no snapshots or anything running at those times. It seems like the fewer VMs there are running on any given Host the shorter time is takes for the CPU to get down to normal. From start to end it takes about 10-15 minutes and then everything cools down again.

Any ideas?

/Peter

 

 

 

Reply
0 Kudos
1 Solution

Accepted Solutions
PeterRaschLager
Enthusiast
Enthusiast
Jump to solution

We have found the root cause: Microfocus uses ZCC to refresh all the servers it knows of and did so every three hours. This has been set to manuel now and the 100% CPU and the data "storm" on the Nimble boxes have subsided.

 

/Peter

 

View solution in original post

5 Replies
depping
Leadership
Leadership
Jump to solution

And you don't have scheduled tasks in all VMs running? Maybe something like anti-virus within the VMs?

Reply
0 Kudos
PeterRaschLager
Enthusiast
Enthusiast
Jump to solution

No.

I see spikes on the NImble boxes at these times on the replication graphs but I cannot says whether this is the cause or effect.

We use Cisco AMP for anti-virus but I have not found any tasks.

 

/Peter

 

Reply
0 Kudos
depping
Leadership
Leadership
Jump to solution

I would suggest running esxtop around the time this happens, if you can't catch it, just run it in batch mode. this will at least show you if the VMs are causing this or not. Could be various reasons, replication/snapshotting could be one of them.

Reply
0 Kudos
PeterRaschLager
Enthusiast
Enthusiast
Jump to solution

We have found the root cause: Microfocus uses ZCC to refresh all the servers it knows of and did so every three hours. This has been set to manuel now and the 100% CPU and the data "storm" on the Nimble boxes have subsided.

 

/Peter

 

depping
Leadership
Leadership
Jump to solution

good find, thanks for looping back!

Reply
0 Kudos