vSwitch Connection Errors

[obrienim]

# esxcfg-vswitch -l

Switch Name Num Ports Used Ports Configured Ports Uplinks

Listing failed for vswitch: vSwitch0, Error: No child key with name=vSwitch0 at parent: /net/vswitch

vSwitch0 32 3 Switch Name Num Ports Used Ports Configured Ports Uplinks

Listing failed for vswitch: vSwitch1, Error: No child key with name=vSwitch1 at parent: /net/vswitch

vSwitch1 16 3 Switch Name Num Ports Used Ports Configured Ports Uplinks

Listing failed for vswitch: vSwitch2, Error: No child key with name=vSwitch2 at parent: /net/vswitch

vSwitch2 16 3 #

I get a similar error message if I try to make any changes via the vi client to the vSwitches.

Added a new guest to the system and tried to connect to Windows Updates with it. Showed no network cable plugged in. As far as I can tell the system appears to be working for the service console on vSwitch1 since I can use the client and telnet to it.

Don't really want to but I can remove all of the network configuration via an iLO connection and recreate it. I am hoping I don't need to go that route, because at this point I ma confused as to why it would stop working and go all crazy.

It was added to a cluster in virtual center and we have tested moving servers off it and that all appeared to work fine then.

strange....

# esxcfg-vswif -l

Name Port Group IP Address Netmask Broadcast Enabled DHCP

#

Doesn't show my vswif for the service console...

ANy ideas?

[obrienim]

Well unless it stored someplace else my server just chewed itself up. I just check /etc/vmware/esx.comf and the size is 0 bytes.

Compared him to another server and needless to say he is missing a few things in the config.

Not sure what to try at this point.

[Chamon]

If you do an ifconfig does the vswif0 show?

How long ago was this Host installed?

Does it have any VMs on it now? If it is a new server you may want to try a repair install or rebuild it. That is strange that it doesn't even see the service console. You said that you can Vmotion VMs to it?

Maybe some of the Gurus have a better answer.

[obrienim]

ifconfig does show vswif0

It was the original ESX host here, so it was at least 8 months ago.

The only VMS on it were ones that I moved back via DRS to test that, ut they had the nics disable so they didn't conflict with some other services. Otherwise there is nothing running production on the server.

We have used vMotion to take servers on and off it no problem. I even created a new server today on it from the VI Client, and it is reporting to Virtual Center.

Rolling over the problems in my head it almost seems that the running configuration is fine, but the config files have been wiped...

[Chamon]

The vSwitch that has the VMs on it have at least on pNIC attached to it?

The VM you are trying to update has had the vNIC attached to the correct vSwitch that has access to the outside.

[Chamon]

Does ifconfig show vmnic0 and vmnic1 upto however many pNICs that you have?

[obrienim]

Took the esxcfg-vswitch info off the mirrored server in the cluster.

Switch Name Num Ports Used Ports Configured Ports Uplinks

vSwitch0 32 6 32 vmnic1,vmnic0

PortGroup Name Internal ID VLAN ID Used Ports Uplinks

Port Group 1 portgroup6 3 0 vmnic0,vmnic1

Port Group 2 portgroup5 335 0 vmnic0,vmnic1

Port Group 3 portgroup4 334 0 vmnic0,vmnic1

Port Group 4 portgroup3 0 3 vmnic0,vmnic1

Switch Name Num Ports Used Ports Configured Ports Uplinks

vSwitch1 16 3 16 vmnic2

PortGroup Name Internal ID VLAN ID Used Ports Uplinks

Service Console portgroup13 0 1 vmnic2

Switch Name Num Ports Used Ports Configured Ports Uplinks

vSwitch2 16 3 16 vmnic3

PortGroup Name Internal ID VLAN ID Used Ports Uplinks

VMkernel portgroup10 0 1 vmnic3

Port Group 4 on vSwitch0 would be the place I ma trying to add a server, the other port groups are not used yet as they match up to physical servers that will be moved into the cluster.

On the switch side of the world it is configred like this for nic0 and nic1, checked the interfaces and they are up on the switch side of the world.

interface Port-channel11

description VMW1_Channel

switchport

switchport trunk encapsulation dot1q

switchport trunk native vlan 339

switchport trunk allowed vlan 3,334,335,339

switchport mode trunk

switchport nonegotiate

spanning-tree portfast trunk

interface GigabitEthernet3/23

description VMW1-vmnic0

switchport trunk encapsulation dot1q

switchport trunk native vlan 339

switchport trunk allowed vlan 3,334,335,339

switchport mode trunk

switchport nonegotiate

channel-group 11 mode on

interface GigabitEthernet4/23

description VMW1-vmnic1

switchport trunk encapsulation dot1q

switchport trunk native vlan 339

switchport trunk allowed vlan 3,334,335,339

switchport mode trunk

switchport nonegotiate

channel-group 11 mode on

[RenaudL]

Yep, you have an empty esx.conf and there's not much you can do about it. If you have a backup of this file, you may use it, otherwise I don't have a lot of solutions to offer you.

What were you precisely doing when your esx.conf disappeared?

[obrienim]

Yep, you have an empty esx.conf and there's not much you can do about it. If you have a backup of this file, you may use it, otherwise I don't have a lot of solutions to offer you.

What were you precisely doing when your esx.conf disappeared?

Don't know that I was doing anything. I found the problem when I couldn't get a new sever to connect to the network.

The only thing I can think of that i have done recently on that server was install and try out the VISBU on Friday. Biggest problem is that the time stamp on the esx.conf file shows today at 11:36 local which would have been about the time I tried to make a confgiuration change in the Port Groups on the server.

[obrienim]

Does ifconfig show vmnic0 and vmnic1 upto however many pNICs that you have?

Shows all 4 pnics and the vswif

[RenaudL]

Hmmm, this is a long shot and I don't really know its internals, but VISBU may have somehow corrupted your esx.conf.

Modifying/reading esx.conf is a risky operation: it should be locked while doing so and esxcfg-* tools are very sensitive to the integrity of this file. If a piece of configuration looks funny, we'd rather discard it than use it.

Is there a /etc/vmware/esx.conf.LOCK?

[obrienim]

Hmmm, this is a long shot and I don't really know its internals, but VISBU may have somehow corrupted your esx.conf.

Modifying/reading esx.conf is a risky operation: it should be locked while doing so and esxcfg-* tools are very sensitive to the integrity of this file. If a piece of configuration looks funny, we'd rather discard it than use it.

Is there a /etc/vmware/esx.conf.LOCK?

No LOCk files, but there is an esx_checksum.conf that has a date stamp of 14:18 (local time) on the 29th of November.

[RenaudL]

Something fishy definitively happened between Thursday and this morning. I'm afraid I don't have any miracle solution to offer you: you'll have to reset the configuration by hand 😕

IMHO, it is always a good idea to backup esx.conf along with your VMs in order to be sure this situation never happens. Hopefully ESX 3.5 will be way more robust on this side.

[obrienim]

Something fishy definitively happened between Thursday and this morning. I'm afraid I don't have any miracle solution to offer you: you'll have to reset the configuration by hand 😕

IMHO, it is always a good idea to backup esx.conf along with your VMs in order to be sure this situation never happens. Hopefully ESX 3.5 will be way more robust on this side.

Only been working with VMware for a couple of weeks, so I am not really sure what blew up the config.

I know it is a good idea to keep backups of the systems files. I just need to figure out a good way to do that.

I did fix the server though.

I generated a vmsupport report (trying to have it ready for the request from them) and when I tried to extract it on my pc it got pretty grumpy with me about file paths and names (guess windows doesn't like those linux fs things). I saw /var/log/oldconf and it had an esx.conf file in there from the same time and day and the checksum file.

Copied the broken one to a new name and copied in the most recent one from oldconf and everything looks good now.

[RenaudL]

Copied the broken one to a new name and copied in the most recent one from oldconf and everything looks good now.

OK, good to know it exists, I wasn't aware of this file myself

[obrienim]

Well scrath the getting it fixed part.

I have the esx.conf back and with datat in it but when I try to do anything from the VI Client it tells me "The item or object refered too could not be found."