1, query the difference time between each node and record, (within 5mins, more than 5mins)
2. Check whether there is a low-status virtual machine on the CNA node (ps aux | grep -w D) Execute this command to find more than one echo, which proves that there is a low-status virtual machine and you need to contact R&D.
3. Configure the clock source (an unused external_OM plane IP address) on the FusionSphere openstack OM interface.
Perform forced synchronization within 5 minutes (you need to configure this IP address to perform forced synchronization on the FusionCOmpute portal interface). The manually modified time exceeds 5mins within 1min of the clock source time.
5. Execute ntpq -p. Check whether the configured clock source exists at the front. If there is a successful configuration, re-find the cause of the problem if the configuration fails.
