ESX Host isolated from network

[ReggieSmith]

I have a host that is giving an error in the console that basically states "Host is isolated from network. either repair network or chage agent startup using -noiso"

How would I do this?

[mprigge]

That usually would indicate that you don't have connectivity to your default gateway or that you don't have a valid service console interface. Either one of those ring true?

[ReggieSmith]

There is supposedly a way to start the network agent on an ESX host to isolate it. I am trying to find out how to adjust that.

All Network adapters are showing UP.

[mprigge]

Ah, okay. So you want it to be isolated? Okay. Little more info is needed here. What are the circumstances that got you where you are? Is this host off by itself on an isolated network? Is it part of a cluster? Are you trying to run HA?

I think the agent it's referring to is probably the Legato HA agent which really shouldn't be run in isolated mode unless you have a really good reason to do so, but I might be wrong - thus the questions.

[ReggieSmith]

No. I DO NOT want it that way but the error I am getting from the console says that it IS. I want to change it using the -noiso option that the error message refers to but I don't know how.

What got me here? good question. all of the suddden I was unable to power on a VM on this host. kept getting an error through VC that "the host could not be contacted." After that VC showed it as "Not Responding"

I tried to putty to the console and couldn't so iLO'd to it and looked at the console. that's when I started noticing the error.

[mprigge]

Gotcha. Okay. I'm going to make the assumption that this host was part of a cluster in which HA was enabled (which makes the error you got make more sense but isn't critical to the problem). If you can't PuTTY into the box and lost VC connectivity, that's your larger problem. The error your seeing there is probably more of an effect/symptom than anything else. So even if we did figure out where that switch goes (probably in the LegatoAAM startup script somewhere), it wouldn't make you feel any better because the host would probably still be unavailable. You're probably going to want to start looking at generalized network issues (of which there could be a whole bunch). Anything from a bad cable, NIC, switch port configuration, unavailable default gateway (probably not that if the other hosts are working), or whatever else can go wrong with your network. Alternatively you could also have something wrong with your ESX host's service console configuration, but I tend to doubt that with the little information I have (assuming no one changed the SC configuration recently).

[ReggieSmith]

All the network settings are correct and the physical (external) network has been verified. I cannot ping out from ANY NIC in this system that is controlled by VMWare. (The iLO connection to the server works fine.) My backup and iSCSI nics, which are on separate networks, cannot ping their gateways either. I have 3 physical NICS for SC and machine traffic and all 3 of those are running over different paths when they leave the box. I doubt that every connection, except the iLO, would die all at once.

I just need to find out how to address the error I am getting about the NODE being isolated from the network and how to use that -noiso option.

[ReggieSmith]

Never found an answer to the location of the "-noiso" option that my error was pointing to so I had decided to simply delete and rebuild my vSwitch.

Since vSwitch0 already existed, I had to delete it 1st. Got an error that the Service Console port group was still in use. I ran "esxcfg-vswitch" to get the option to delete it and saw the "-r" option. (Restore all virtual switches from configuration file.) So I ran "esxcfg-vswitch -r" and my connectivity was restored.

I guess this was the way to "repair" my network as indicated in the error I was receiving. I would still like to know the location and use of the mysterious "-noiso" option. If anyone can find the documentation on it i will give them credit for it.

[xxzab]

I had this same problem it started after I tried to add and new NIC to the vSwitch0 and VCC crashed when doing it. Had to run "[b]esxcfg-vswitch -U vmnic1 vSwitch0[/b]" that gave me "[b]Node no longer ISOLATED, will RESTART[/b]" and the network started to work again after.

[dstrik]

Can anyone answer how to use the -nosio option?

I know how I managed to isolate one of my HA cluster nodes, it would be nice to know how use -noiso to fix the problem!

[dstauffer]

The -noiso is an option for the /opt/LGTOaam512/bin/ft_startup script. The ftp_startup script is called from the /etc/init.d/LGTOAAM51_vmware script. VMware's HA is a Legato product. Legato has documentation which in this case explains the -noiso option. The documentation is available at:

http://web1.legato.com/cgi-bin/catalog?sf=Releases&level=29-1#29-1

If you are running ESX 3.0.1 you are using Legato Automated Availability Manager 5.1.2.