Virtual Swith memory

[AlexEv123]

Something strange happens in my ESXI 7. I have from provider a lot of IPs, IPs assign my provider to two VLAN, two my virtual switch, two different network

But unfortunately I can use only 36 IPs in both switch, If I start 37 Virtual Machine with 37 IP, my last 37 machine can not see provider gateway and can not access to inet. However I see MAC-addr of last 37 machine in my VirtSwitch. It this case last machine is Redis, it was working before I start U18. If I will shutdown U18 than Redis will be see gateway.

What a reason of this restriction? I think there is not enough memory for my Virtual Switch, isn't it? Or something other restriction? How I can overcome this strange restriction?

[scott28tt]

@AlexEv123 

It may be helpful to show the properties of your Outside-VLAN4001 port group.

[depping]

there's no such restriction, in terms of memory. As already asked, can you show the portgroup configuration? Also, is there an error message anywhere when you go beyond the number that works (events? / logs?)

you can go to the commmandline of a host to easily check the current number of ports it is using and how many can be used for the vswitch:

esxcfg-vswitch -l

[AlexEv123]

1. Thank you, @scott28tt, this is my port configuration

[AlexEv123]

Thank you, @depping, this is response of available ports, in log I can see only normal message about my operation, there is no criminal or interesting message at all

[depping]

You definitely should not be hitting this problem based on the configuration of the switch. It can't have anything to do with memory. You sure the network side is actually correctly configured? (subnet etc? was there a miscalculation on that side maybe?) 

[scott28tt]

@AlexEv123 

At the times when the VM OS cannot get to the outside world, what connectivity does that OS have in relation to other VM OSes or anything else?

[AlexEv123]

hm....

Last 37 machine see all other machines perfectly but can not see gateway and can not have access to inet.

Most my client's machine is simple Ubuntu, I have access to that machines and can not see any issues in configuration. All configuration network configuration in Ubuntu is simple and similar.

Is it a something datacenter restriction?

[a_p_]

From what I understand it my be a restriction on the provider's side. Something like a limitation of allowed MAC addresses on the uplink port!?

André

[AlexEv123]

maybe, but I need hard and strong prove this restriction, because datacenter provider newer answer to tickets correctly, usually ticket to provider it is a dead unsuccessfully end of attempt to solving any problem.

[AlexEv123]

and finally, after experiments I understand condition where my issue is happens

in my case VM working fine before I will shutdown it

since shutdown moment if I boot machine not immediately, maybe wait a couple of hours, VM has no access to inet (but only with current IP!!)

if I change old IP to another than VM will be working fine, or if booting will be without big timeout

I ask yesterday datacenter provider, but not receive concrete answer

does anybody understanding what happens in my case and how it's possible to save current IP addr after VM stored long time in inactive state?