Shulou(Shulou.com)06/02 Report--

If you create multiple vSAN clusters in the same vSphere data center, the ESXi host of each vSAN cluster should be a newly installed operating system. If you remove ESXi hosts from one vSAN cluster and create a new vSAN cluster in the same data center, only one vSAN storage is displayed and the capacity size of the vSAN storage is not displayed correctly. This failure occurred recently during testing with vSphere 6.7.0 U3, so let's review the lab now and hope it will be helpful to other readers. In this experiment, the ESXi version number is 6.7.0-14320388 Server version number is 6.7.0-14368073, which is the version of vSphere 6.7.0 U3.

1 problem phenomenon

At present, there is an experimental environment for vSAN, four ESXi hosts, the management address is 172.18.96.41fu4Magi vsan, the address is 192.168.0.41 4th Magi vsan, and a 10 Gigabit switch is specially used. The management address is connected to another gigabit switch. Each machine has 64G memory, each machine has a disk group, 1 SSD and 2 2TB. This vSAN is working normally. Normally, the capacity of vSAN storage is 14.55TB.

Then in this vCenter, a new cluster, 2-node vSAN, is added. Each ESXi host has 96G memory, and the two host vSAN traffic network cards are directly connected by optical fiber (10 Gigabit network cards). Each unit has 1 disk group and 1 SSD,3 block 2TB. Witness that the host is in a separate ESXi host. Normally, the capacity of this vSAN storage is 10.92TB.

Since the addition of this vSAN cluster, in vSphere Client, whether you browse the original vSAN cluster or the newly added vSAN cluster, the only vSAN storage displayed is the one that was added later (capacity is 10.92TB). As shown in figures 1, 2, and 3.

Figure 1 viewing storage in the data center

Figure 22 Node directly connected to View Storage

Figure 3 original vSAN storage

The original vSAN cluster disk management and vSAN monitoring information are normal, as shown in figure 4.

Figure 4 original vSAN disk management

The second vSAN cluster disk management also appears normal, as shown in figure 5.

Figure 5 disk management of the second vSAN cluster

Log in to one of the hosts in the first vSAN cluster using SSH, execute esxcli vsan cluster unicastagent list, and see that the unicast message is normal, as shown in figure 6.

Figure 6

Log in to a host in the second vSAN cluster, for example, 172.18.96.45, and the information is normal, as shown in figure 7.

Figure 7

Figure 8 first vSAN Cluster 2 vSAN Monitoring

Figure 9 second vSAN cluster

After trying to remove the second vSAN cluster, the first vSAN cluster vSAN storage displays as 0 (as shown in figure 10) and returns to normal after disabling HA and then enabling it.

Figure 10 vSAN is stored as 0

Figure 11 normal after reconfiguring HA

Then create a new vSAN cluster in the data center again, and the failure occurs again after adding 2-node hosts.

2 create multiple data centers

Because the previous experiment was to create multiple clusters in the same data center. Next, two data centers are created, each with a vSAN cluster. The same lab host is still used when adding the vSAN cluster. However, this lab found that when multiple data centers were created in the same vCenter and each data center had a vSAN cluster, the vSAN storage in each data center appeared normally. As shown in figure 12, there are three data centers in the figure, each with one vSAN cluster, and three vSAN storage can be displayed at the vCenter level.

Figure 12 multiple data centers and multiple clusters appear normal

Since multiple data centers, multiple clusters, and vSAN storage appear normal, multiple clusters in the same data center should also be possible. Because similar experiments have been done before. It has been done before when vSphere 6.5 and vSphere 6.7, but this time vSphere 6.7.0 U3 went wrong and thought it was a new version of bug.

Later, I thought that these computers were initially in a vSAN cluster, which was initially a vSAN cluster of 12 hosts, but finally removed 8 hosts from the cluster, shrinking the vSAN cluster to 4. Is it possible that the eight ESXi hosts removed from the cluster did not reinstall the system?

Then I kept only the vSAN cluster of 172.18.96.41 to 172.18.96.44, and all the other hosts were reinstalled. After reinstallation, two clusters were created in a data center, each adding different ESXi hosts and configuring vSAN. After this lab, it was found that the experiment was successful, as shown in figure 13.

Fig. 13 successful experiment

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.