Solved

Does Apps management (Calm) support ESXi in PrismCentral 5.8?

  • 1 August 2018
  • 7 replies
  • 2660 views

I deployed a Nutanix Cluster (AOS version 5.8) with ESXi 6.5. And I deployed a Prism Central 5.8 on it. There're some articles on internet, saying App Management supports ESXi, starting from PrismCentral version 5.7. Is it true?

In my case, I can't find Self Service & Apps from the pull-down list, when I click question mark icon in the main menu on PrismCentral 5.8. I did see "PC Scale out", "Enable app management" and "Microsegmentation", but not "Self Service & Apps".

When I click "Enable app management", I see the following error.
  • Precheck failures:
Host Cluster Registered
Please check that the cluster on which this Prism Central is hosted is registered to it.

I had already registered my Nutanix cluster into this PrismCentral. It looks to me that App management or self service doesn't support ESXi, correct? Or is there anything I missed?
icon

Best answer by calm_mark 2 August 2018, 19:37

View original

This topic has been closed for comments

7 replies

Userlevel 2
Badge +2
Hello @weiyi ,

Yes, Calm inside of Prism Central has supported ESX workloads on and off Nutanix clusters since release 5.7 back in May, right before the Nutanix .Next US conference.

"Enable App Management" is Calm!

This paragraph has forward looking statements; verify Nutanix has shipped these features before making any additional purchasing decisions. You will see the Calm team drive the current "Self Service & Apps" experience into Calm over time, starting with the next major release targeted as Prism Central 5.9 before the end of the calendar year. You will see a continued convergence of Calm into Prism and alignment with other products, such as Beam for application total cost of ownership.

Now, onto your question: it's hard to troubleshoot fully without more information, but it seems as if you have fresh Nutanix cluster with ESX, but something does not add up because Prism Central is not fully registered to PE.

In the Nutanix Support Portal, please review the Documentation > Software > Prism Central 5.8 Guide, I've pointed you to the procedure to unregister Prism Central and destroy the PC VM(s) so you can try to launch Prism Central again properly registered with your cluster. You should perform a NCC check to insure your cluster is in the best state before Prism Central launching again.

Furthermore, you should insure memory hot add is enabled for Prism Central VMs before "Enabling App Management," see: Enabling Calm on a Nutanix VMWare cluster, I believe we do this automatically on AHV.

I would strongly suggest you call Nutanix Support before you do any of this so they can help you along the way, answer questions, and troubleshoot further.

Thanks for trying Calm!
--Mark
Mark,
I unregistered the PE from PC as described by the document. And I removed the PC in vCenter, with the action of "Delete from disk". After that, I upgrade PE from 5.8 to 5.8.0.1.

And then deployed a PC of version 5.8.0.1. But after PC was deployed, PE was not registered into PC automatically. But as I remember, PE should be registered into PC automatically after PC deployment before, correct?

After PC deployment, I registered PE into PC manually. The App management still doesn't work. It shows the error of "Host Cluster Registered: Please check that the cluster on which this Prism Central is hosted is registered to it.", when I launch "Enable app management".

I ran the "ncc health_checks run_all" on both PE and PC. Both of them looks good. Here's the output:

Prism Element:

Detailed information for ondisk_dedup_enabled_check:
Node 172.29.0.8:
INFO: On Disk Dedup is disabled by Stargate.
Refer to KB 1851 (http://portal.nutanix.com/kb/1851) for details on ondisk_dedup_enabled_check or Recheck with: ncc health_checks stargate_checks ondisk_dedup_enabled_check

Detailed information for dns_server_check:
Node 172.29.0.8:
INFO: Only one name server is configured
Node 172.29.0.7:
INFO: Only one name server is configured
Node 172.29.0.13:
INFO: Only one name server is configured
Refer to KB 3005 (http://portal.nutanix.com/kb/3005) for details on dns_server_check or Recheck with: ncc health_checks system_checks dns_server_check --cvm_list=172.29.0.8,172.29.0.7,172.29.0.13

Detailed information for ipmi_sensor_threshold_check:
Node 172.29.0.8:
FAIL: Fan Speed Low
ERR : Invalid sensor info received
Refer to KB 1524 (http://portal.nutanix.com/kb/1524) for details on ipmi_sensor_threshold_check or Recheck with: ncc health_checks hardware_checks ipmi_checks ipmi_sensor_threshold_check --cvm_list=172.29.0.8
+---------------+
| State | Count |
+---------------+
| Pass | 200 |
| Info | 2 |
| Error | 1 |
| Total | 203 |
+---------------+
Plugin output written to /home/nutanix/data/logs/ncc-output-latest.log

PrismCentral:

Detailed information for dns_server_check:
Node 172.29.17.185:
INFO: Only one name server is configured
Refer to KB 3005 (http://portal.nutanix.com/kb/3005) for details on dns_server_check or Recheck with: ncc health_checks system_checks dns_server_check --cvm_list=172.29.17.185
+---------------+
| State | Count |
+---------------+
| Pass | 24 |
| Info | 1 |
| Total | 25 |
+---------------+
Plugin output written to /home/nutanix/data/logs/ncc-output-latest.log

And also I ran the command of "ncli multicluster get-cluster-state" on both PE and PC. PE looks good. But PC shows "Is Multicluster" as false. Is it correct? What else I could check?

Prism Element:

Cluster Id : 99e63133-e634-4836-98da-3e69b519cc40
Cluster Name : Unnamed
Is Multicluster : true
Controller VM IP Addre... : [172.29.17.185]
External IP Address :
Marked for Removal : false
Remote Connection Exists : true

Prism Central:

Cluster Id : 00057053-e34b-795d-41cc-0894ef0339f0
Cluster Name : TestCluster
Is Multicluster : false
Controller VM IP Addre... : [172.29.0.7, 172.29.0.8, 172.29.0.13]
External IP Address :
Marked for Removal : false
Remote Connection Exists : true
By the way, it seems that everything works fine, other than "Enable app management". I can see PE cluster in PC. I can see all VMs, containers, alerts etc in PC.
Userlevel 2
Badge +2
@weiyi Thanks for providing that information: nothing stands out immediately to me except external IPs not being defined.

Multicluster on PC probably indicates that you deployed PC as a single instance, not with the scale-out option.

Please check on two items:

1) On PE, can you define an Data Service IP? Calm uses this to store logs on Acropolis storage on the cluster.

2) Could you check that memory hot add is enabled for Prism Central VMs before "Enabling App Management," see: Enabling Calm on a Nutanix VMWare cluster?

If you have already done both, then I'd recommend calling Nutanix Support at this time.
--Mark
Yes, I've configured the "ISCSI DATA SERVICES IP" in Prism Element. And also I checked my vCenter 6.5 that the value of "mem.hotadd" is "TRUE".

Do you think the external IP need to be set? 5.8.0.1 is a community version, but not a public one. Can I call Nutanix support?
Userlevel 2
Badge +2
@weiyi Good, it appears you've followed the documentation procedures for set up.

Aside from going into troubleshooting the logs on the CVMs on a deeper level, we've gone past my troubleshooting abilities and we'll have to see if anyone else in the community can help. I'm sorry, I don't have the ability to reproduce your setup to help you further.

You are correct: community support is the best we can offer you unless you're willing to change this to a supported cluster.

Update: I've conferred with my colleagues and there is an issue with the latest VMWare ESX 6.5 and activation of Calm in Prism Central: it was identified in the Prism Central 5.8 Release Notes: Known Issues section.

The Calm team has a fix in code review and then it will be ready for release planning, so I can't give you a firm date yet. In the meantime, please see this Knowledge Base article = https://portal.nutanix.com/kb/5553 for a workaround with your current installation.

Our apologies for the inconvenience!
--Mark
The KB works https://portal.nutanix.com/kb/5553

I searched the error message on google. But I didn't find the KB or any useful information. It works now. Thank you very much!