Migrated Linux machine won't start
-
I'm fairly new to XCP-ng and am working on a project to migrate some virtual machines from my older Vmware 6.7 host to a new XCP-NG build. Its a single stand alone server and is basically default. I have successfully migrated 5 Windows VM's without issue, knock on wood. However, I have a linux based virtual machine that runs my phone system. It seems to import to XCP just fine, and it even seems to start fine. It starts to boot up and shows no errors but for some reason at about 90 seconds the VM just shuts off. No error, no exit anything. Best I can describe it, is it seems like the virutal power plug gets pulled on it. I can't find any logs that would indicate the vm is being shut off. It's just off.. Any ideas?
-
@ItsAlwaysDNS Have you checked your DNS?
No but seriously is the VMWare guest tool installed on this system? What version of linux is it using?
-
When the VM is running on the Vmware host, it reports the tools are running. Looks like its running CentOS 6 (32bit).
-
Check to see if the VM is running in PV mode -- https://xcp-ng.org/blog/2022/01/17/removing-support-for-32-bit-pv-guests/
-
@Danp On the advanced tab of the VM, its set to HVM.
-
@ItsAlwaysDNS said in Migrated Linux machine won't start:
When the VM is running on the Vmware host, it reports the tools are running. Looks like its running CentOS 6 (32bit).
CentOS 6 is very out of date, but I can relate to keeping something around longer than it really should be.
Can you make a copy of the VM on your ESXi host, remove the guest tools from the VM and try to migrate the VM without the guest tools?
-
I can try to give that a go. Is there any particular reason as to why we think this will work? I'm totally fine with the thought of "its worth a try". I'm just curious if there is some known compatibility issue that is leading you to suggest trying this as a solution.
-
@ItsAlwaysDNS I still have some old Debian systems running just fine. XCP 8.2.1, HVM, Debian 7, upgraded kernel 3.16 (32 bit). They were cloned from hardware systems and just added the XCP agent. I have run even older systems (same age as CentOS 6) and they work correctly.
You can look at this old CentOS 6 post.
-
@ItsAlwaysDNS said in Migrated Linux machine won't start:
I can try to give that a go. Is there any particular reason as to why we think this will work? I'm totally fine with the thought of "its worth a try". I'm just curious if there is some known compatibility issue that is leading you to suggest trying this as a solution.
Normally these older systems were required to use the ESXI drivers, these drivers are also normally very out of date and cause issues.
Most of the time, when these issues occur, it's because of incompatible drivers. Cloning the VM and removing the drivers from the clone and then trying to migrate the VM is the safest way to confirm the issue is with the drivers without impacting your production system.
