VMware-Forum

Hi,

I crashed my ESXi network adapter/ configuration.
I got ESXi 5.1 installed and accessing (when functioning) with vSphere Client 5.1.

The physical adapter was set to "handle out" speed and duplexity. Well it was handled to 100m but I wanted to see 1000m.
Thats why I changed manually to "1000m fully duplex". After saving and some seconds of load I got an errormessage (which I

cannot remeber exactly). It was about an timeout.
The messages raises a second time and after that I lost connection. Well I am not absolutely sure if it was the timeout of

my vSphere Client after losing connection earlier than I recognized.
I went to console of my ESXi and tried some built-in functionalities of ESXi console:
restarting the management network, restarting the ESXi and restoring network configuration.
The first both did not seem to have any effect.
The lasts'effect was that I am not able any more to configure network via console (shown greyed).
Syslog says: "Unable to restor VMkernel default gateway (172.17.0.1): unable to set VMkernel default gateway address.

Please verify your IP settings and try again".
Which I cannt do since the function is shown greyed.

I thinking about usin another built-in functionality: "Reset System Configuration".
But I am not sure what effect this step will have.
I am afraid of losing my storage-mountings and therefore my assigned virtual machines.
Losing network settings and such things werent that problem I think.

I remeber having the following network configuration before crashes:
2 internal networks (LAN0 and LAN1) including a virtual switch each
with each 1 physical adapter (at this moment just the first one (LAN0) is physically connected to the network)
a management network in both internal lans (I wanted to change managing network from LAN 1 to LAN0)
some virtual machines in both networks (was going to switch them all to LAN0)

physically my server has got 2 LAN ports and 1 (I guess) console-port (its a RJ-45 with a wrench as icon).
The console I am connecting with is accessed via VGA and PS/2 directly.

So there are several questions I have:
- will I get it work with restoring system configuration?
- do I have to reinstall the whole server?
- are there some console commands which I can use to restore my network (adapter)?
- if yes: which, where can I find a documentation about them?
- if yes and it is dangerous to fire some commands, any advise how exactly to fire them would be nice

thanks in advance.

To see 1Gbit/full dupplex you need the right switch and proper cable also. Not alway auto negotiation will work so you need to pinpoint 1GB Full on switch side too.

Reboot your host and try to restore Management on VSS of console doesnt work any more.

Regards
Joerg

If the initial setting of your network adapter was "auto", and your environment caused it to run at 100M, then the following procedure should help:

- Login locally at the server's console (enable ESXi shell in DCUI's Troubleshooting Options, then alt-F1, login root)
- run "esxcfg-vmknic -l" to determine the management interface's vmkernel portgroup (normally "management network" on vmk0)
- run "esxcfg-vswitch -l" to determine the vmnicX associated with this portgroup
- run "esxcli network nic set -n vmnicX -a" to set its "speed and duplexity settings to autonegotiate"

To answer your questions:

- will I get it work with restoring system configuration?

Most definitely yes, but you will need to find a way to restore a previosly stored system configuration. In case you meant "resetting system configuration", the answer is still yes. You will lose the registration of your VMs, but NOT the VMs themselves. Simply re-register them after reboot.

- do I have to reinstall the whole server?

No.

- are there some console commands which I can use to restore my network (adapter)?

See above.

- if yes: which, where can I find a documentation about them?

In vSphere CLI documentation. Sorry, I don't remember the link. It has been mentioned on this forum already, otherwise, pls. use your favorite search engine.

- if yes and it is dangerous to fire some commands, any advise how exactly to fire them would be nice

Not really dangerous; you can always reset the system configuration...
But I hope the description of steps outlined above will be sufficient.

hey thanks four your answer

@irix: system reboot still done, built-in function for recover also (as i described)
I dont understand what youre meaning with "on VSS of console doesnt work any more. "

@JustMe: what for do I need to run the first both commands? Do I have to replace vmnicX (I strongly guess) with the specific name?
what do I Have to replace by the first command? (is it "network"?)

I will give it a try later.

If you still remember the right vmnicX, then you won't need the first two commands. They're just meant to ultimately determine the correct X value for vmnicX.

As for the typing:
Just type the first two commands as they are depicted between the apostrophes. The output will need some interpretation, of course, but cannot be predicted by me, just guessed.
The output will look similar to this:


The first command shows that my vmk0 is bound to portgroupt "Management Network", and the second command reveals that "Management Network" is backed by vmnic3.

As a consequence, I would have to type

to achieve the desired result.

hey great
after prompting "esxcli network nic set -n vmnic0 -a" nothing happened (seem to finish directly and without return or message).

this is what esxcfg-vswitch -l results:


restarted Server but no difference.
what could I do now?
thanks

vSwitch0 doesnt have any pNIC assigned. So add one for testing and see if you can configure a MGMT from DCUI.

Regards
Joerg

What do you get from ?

Could you please put the output into a "code" context for better readability? Thanks.

If all else fails, you might need to bite the bullet, and "reset configuration"...

hey

esxcli network nic list creates the following result:




i also fired

to achieve what irix suggested (assign nic to vswitch)

got the following error:
Failed to add uplink vmnic0 to Switch0, Error: Sysinfo error on operation returned status : Busy. Please see the VMkernel log for detailed error information.

which is:
NetTeam: 558: Failed to register uplink vmnic0: Busy



Well, any further ideas before trying reset?

thanks

You could try to first Unlink vmnic0 from vSwitch1...

But anyway, there could be more parameters corrupted. With both vmnic0 and vmnic1 linked to your Management Network vSwitch, and vmnic0 detected as down, ESXi should simply route all traffic through vmnic1 in an otherwise standard configuration, I guess.

As this does not seem to happen, you'd better drop what you have right now, and "reset system configuration", to start all over again.

hey

did change vmnic but nothing happens.

Now I restarted whole configuration and it worked.
I just lost assigned VMs.
Even all storages are still available.

Thanks for your help, I learned much.