Hello
I have astrange behavior.
My ESX 3.5 rebooted and I don't have anymore access by network to my console.
All the VM started and are working fine.
If I log to the "physical" console, I can see the correct IP config, ping myself ... everything is fine but no network comunication with any other host
Any insights ?
Thx
We had this issue at our place once and it turned out it was a dodgy network card in the blade as opposed to anything to do with ESX. Have you ran any diags to see if it is the blade ? Ours was showing as connected until we ran diags and rebooted the blade and then it started to report errors about the physical NICS.
Did you check the light/LED on the NIC and on the physical switch port? If this is a managed switch, do you see the MAC address logged in to the physical switch port? Any configuration issues on the switch port (port security enabled, spanning tree)? Network cable or connection issue?
Did you check for a duplicate IP address in the network?
André
It's a blade center, so no issues with the cable ..
it already happened a long time ago and the solution was to reinstall from scratch the ESX... if I can avoid this and find the "WHY"
enough space, no issues ( checked with df -h )
Hi..
Is there multiple NICs connected to your vSwitch where your vswif resides?
Im thinking that maybe ESX decided to put your vswif traffic onto another physical NIC within the same vSwitch after reboot, but this physical NIC is missing your management network..
You could try to link /unlink vmnics to this vSwitch from console, to see if this is the issue..
Just thoughts...
/Rubeck
yes I have multiples ones but as it is a blade, the 2 network card are connected on the same "location"
so even it was switched, there is no difference , no ?
If you do "esxcfg-vswitch -l" on the console, you should then see two vmnics connected to your Service Console port group.
Unlink one of them from the vSwitch (example: "esxcfg-vswitch -U vmnic0 vSwitch0")
If it dosn't start to reply then add it back again (example: esxcfg-vswitch -L vmnic0 vSwitch0")
Then unlink the other one (example "esxcfg-vswitch -U vmnic1 vSwitch0").
This approach would force the vswif (and other portgroups configured on the vSwitch) to change the active pNIC.
If it starts to reply, there's config error in what-ever physical switchport the blade is connected to..
Hope it makes sense.. .
/Rubeck
If you do "esxcfg-vswitch -l" on the console, you should then see two vmnics connected to your Service Console port group.
Unlink one of them from the vSwitch (example: "esxcfg-vswitch -U vmnic0 vSwitch0")
no, only one
( Pict 1 )
Now I remove all the config to be sure ...
esxcfg-vswitch -U vmnic0 vSwitch0
( Pict 2 )
othen add again vmnic0 in case off
esxcfg-vswitch -L vmnic0 vSwitch0
( Pict 3 )
trying to ping my gateway or another host on the net, still nothing ...
What i dont understand is that I should be abel at least to ping or do anything at network level ? having a console linked is something required for ESX management no ?
If it dosn't start to reply then add it back again (example: esxcfg-vswitch -L vmnic0 vSwitch0")
Then unlink the other one (example "esxcfg-vswitch -U vmnic1 vSwitch0").
This approach would force the vswif (and other portgroups configured on the vSwitch) to change the active pNIC.
If it starts to reply, there's config error in what-ever physical switchport the blade is connected to..
Should I remove everything, reboot and add it again ?
We had this issue at our place once and it turned out it was a dodgy network card in the blade as opposed to anything to do with ESX. Have you ran any diags to see if it is the blade ? Ours was showing as connected until we ran diags and rebooted the blade and then it started to report errors about the physical NICS.
I had issues to upload the pic t... but that's it ...
After removing all nics and rebooting, I had an issue with the NIC ...
I will investigate, but nothing on ESX side ...
Thx !