Test instance problem while migrating centos 6.8 from virtualbox to AWS

Question

We are migrating from VirtualBox(6.1) virtualization to AWS for MGN testing.
The source server is Centos 6.8.

By installing a replication agent on the source server, it went smoothly with 'ready for testing' status.
However, after the 'launch test instance' action, the test instance created in the 'Test in progress' phase does not operate normally.

Status in Lifecycle is as follows.
- Launch status
  Launched
  First boot: Stopped
- Last test : Completed (Check through Job Id link)

Data replication status is as follows.
- Healthy
- Replication pregress : Initial replication finished

In the instance menu, the test instance status is running, but the status check is 'initialized' for more than 30 minutes, and the instance status check fails.  (System status check is successful.)

Nevertheless, when I look at the instance screenshots, the prompts on the console screen look fine.

After allowing ssh and ping to the security group, ping and ssh connection are attempted from the bastion host, but it fails.

I don't even know if the replication of the whole instance is the problem, or just a part of the instance (eg the network) is the problem.

If you have a similar experience on MGN, please give me a hint.
Thank you.

Answer

Now that we have a clue about the phenomenon, we are asking for specific help by sharing it.

The 'instance status check fail' in the text is because the network card is not normally attached to the migrated cent6.8 instance.

When I check this in console with the 'ifconfig -a' or 'ip a' command
I have confirmed that only the loopback interface exists.

Since the following response was received with the 'lspci' command, it is assumed that the network card on the hardware is installed.
------------------------------------------------------------ ---------------
00:05.0 Ethernet controller: Device 1d0f:ec20

If you receive a detailed response with the 'lspci -v' command, it is as follows.
------------------------------------------------------------ ---------------
00:05.0 Ethernet controller: Device 1d0f:ec20
Physical Slot: 5
Flags: fast devsel
Memory at febf4000 (32-bit, non-prefetchable) [size=16K]
Capabilities: [70] Express Endpoint, MSI 00
Capabilities: [b0] MSI-X: Enable- Count=9 Masked-

Therefore, we presume that it is because the centos 6.8 instance does not have the driver that can recognize the network card provided by AWS.

So, specifically, I am asking for help on how to make AWS ENI aware of a centos 6.8 instance (which has been migrated).

Answer

I had similar issue while migrating CentOS 6.x to AWS via MGN,
In System Logs of the launched EC2 saw message:
 **Bringing up interface eth0:  Device eth0 does not seem to be present, delaying initialization.**
Fixed this by changing EC2 Instance Type from t3.xxxx to t2.xxxxx due to this:

* t3.xxxx is **Nitro System** Hypervisor
* t2.xxxx is **XEN** Hypervisor

Reference:
Nitro instance types will work with RHEL 7.0+/CentOS 7.0+/Oracle Linux 7.0+. This specific limitation does not apply to other instance type families.
https://docs.aws.amazon.com/mgn/latest/ug/Supported-Operating-Systems.html

Test instance problem while migrating centos 6.8 from virtualbox to AWS

Relevant content