Baremetal instance stuck in starting state and timeout after 30 minutes.

[yummydsky]

ISSUE TYPE

• Bug Report

COMPONENT NAME

Baremetal

CLOUDSTACK VERSION

4.19.1.3

CONFIGURATION

OS / ENVIRONMENT

Cloudstack 4.19.1.3 on Ubuntu 22.04LTS

SUMMARY

I use Ubuntu 22.04 LTS as ISO to deploy a bare metal instance, but it seems that bare metal instance stuck in starting state and timeout like issue https://issues.apache.org/jira/browse/CLOUDSTACK-9716

2024-12-02 06:24:44,311 DEBUG [c.c.b.n.BareMetalResourceBase] (DirectAgent-39:ctx-6061c3d0) (logid:47b584a0) still wait for baremetal provision done notification for vm[name:i-2-30-VM], current vm state is Starting

STEPS TO REPRODUCE

EXPECTED RESULTS

The bare metal instance can be deployed os successfully and booted up.

ACTUAL RESULTS

The bare metal stuck in starting state because there is no one to notify and change the vm state to Running

[boring-cyborg[bot]]

Thanks for opening your first issue here! Be sure to follow the issue template!

[DaanHoogland]

thanks @yummydsky , baremetal has not been well maintained for a long time. Have you got other bare-metal instances that are running fine?

[yummydsky]

thanks @yummydsky , baremetal has not been well maintained for a long time. Have you got other bare-metal instances that are running fine?

No, there are no bare-metal instances that can be created properly.

[yummydsky]

@DaanHoogland Is there a PXE server inside the virtual router? I found a document(http://events17.linuxfoundation.org/sites/events/files/slides/Baremetal-v4.1a.pdf) that said we need an additional VR to act as a PXE server. Do you know how I can add a VR as a PXE server?

[DaanHoogland]

@yummydsky I do not know of this , but it seems that adding a pxe server itself to a VR is not a big deal. configuring it properly may vary depending of your infrastructure and requirements.
@rajujith Do you have any experience with this?

[rajujith]

@DaanHoogland, I know this feature, but I don't recall trying it. @harikrishna-patnala may be able to help. However, this feature requires a basic zone based on the document you mentioned? also specific switch model? (http://events17.linuxfoundation.org/sites/events/files/slides/Baremetal-v4.1a.pdf). cc @yummydsky

[yummydsky]

@rajujith
I have no idea about what I need to prepare to give it a try. I assume that there should be a PXE server inside if I choose the network offering that offers the following services.

Supported services
BaremetalPxeService: BaremetalPxeProvider
Dhcp: BaremetalDhcpProvider

I wonder if I can deploy a bare-metal instance with the Advanced Zone because we need to separate the management network and guest network.

If someone can help to provide the details for each step that would be appreciated.

[yummydsky]

I found out there is a check if the NetworkType is not Basic, then BareMetal will report BaremetalDhcpElement can not handle networkoffering: pxe-network
Does this mean we can only use the basic network for the BareMetal machine?
https://github.com/apache/cloudstack/blob/main/plugins/hypervisors/baremetal/src/main/java/com/cloud/baremetal/networkservice/BaremetalDhcpElement.java#L86

[DaanHoogland]

@yummydsky , did you find a resolution to this query?

[yummydsky]

No, so far, I have no resolution to this function.

[DaanHoogland]

not nice, as I said the baremetal plugin has not been well maintained. There is work going on for a “deploy anything” plugin though (see #9752) and I have seen it being used for baremetal. But this is still experimental and will be in v21 the earliest.