HCX Manager high CPU condition
Article ID: 321658
Updated On:
Products
Issue/Introduction
- The performance of the HCX Manager is degraded, and the UI is slow or unresponsive.
- In the vSphere Performance charts, you can observe high CPU utilization for the HCX Manager VM:
- The top command output will display a similar output as shown below:
-
top - 08:35:29 up 137 days, 23:09, 1 user, load average: 7.53, 6.18, 6.05 Tasks: 229 total, 5 running, 224 sleeping, 0 stopped, 0 zombie %Cpu0 : 100.0/0.0 100[|||||||||||||||||||||||||||||||||||||||||||||||||||||] %Cpu1 : 88.2/11.8 100[|||||||||||||||||||||||||||||||||||||||||||||||||||||] %Cpu2 : 100.0/0.0 100[|||||||||||||||||||||||||||||||||||||||||||||||||||||] %Cpu3 : 93.8/6.2 100[|||||||||||||||||||||||||||||||||||||||||||||||||||||] GiB Mem : 28.9/11.7 [ ] GiB Swap: 75.8/4.0 [ ]
-
- From the top command (using Shift + P), the process consuming the most CPU is related to PostgreSQL (psql):
-
18858 postgres 20 0 409.7m 344.9m 87.5 2.9 1338:38 R `- postgres: postgres hybridity 127.0.0.1(55886) SELECT 5175 postgres 20 0 395.5m 330.1m 93.8 2.8 1169:00 R `- postgres: postgres hybridity 127.0.0.1(53234) SELECT 6445 postgres 20 0 391.8m 338.3m 100.0 2.8 546:24.95 R `- postgres: postgres hybridity 127.0.0.1(47594) SELECT 29975 postgres 20 0 361.5m 316.7m 100.0 2.6 8:20.16 R `- postgres: postgres hybridity 127.0.0.1(55916) SELECT
-
Environment
HCX
Cause
In environments with more than 100 ESXi hosts, the large number of objects in the vCenter inventory, such as virtual machines, port groups, and datastores, can lead to high CPU utilization. In addition, large tables can cause the Postgres process to reach 100% CPU utilization.
Resolution
HCX 4.10 includes performance improvements and enhancements. Additionally, ongoing efforts are being made to address this issue, with new performance enhancements planned for future releases.
Workaround:
If you believe you have encountered this issue, please open a support case with Broadcom Support and refer to this KB article.
For more information, see Creating and managing Broadcom support cases.
Additional Information
The CPU condition has been observed for HCX Connector/Cloud manager running 4.4.0 or later.
IMPORTANT: The CPU condition highlighted in this DOC is fully dependent upon the ESXi host scale and may vary depending on the type of deployment and infrastructure conditions.- There will be NO impact on the HCX Interconnect IX appliance.
- There will be NO impact on the migration performance.
- The high CPU condition will NOT impact the Network Extension datapath.
Feedback
