Lossing pings and dropping connection

[mihalina]

Hello,

I recently installed ESXi. I have two NICs. One is for local network and the other one is for public ip-s. The other one is working perfectly but the first one (private ip) is causing me a lot of problems. If I connect to the esxi via vsphere the connection is dropping all the time. I have also ping one VM from switch (cisco 2950) on private ip and the pings were 47/50 successful.
On this server I was also having Xen server and it was working perfectly but I replace it with esx due to my sas driver problem (new controller). The interesting this is if I put auto-negotiate in Vsphere and auto on cisco, the port come up but the problem is the same (connection is dropping).
If I manually put bouth on full duplex the port does not come up and in CLI in switch is saying that the connection is half duplex.

I also checked my MAC addresses and everything is fine.

Thank you very much for your help!

Sorry for my English :smileyhappy:

Miha

[AWo]

Welcome to the forums!

WHat kind of network connectionis having the problem? You talk about full and half duplex, so is ist 100 MBit/s? If so, according to my expierence it is better to have the port settings fixed and not autonegotiate in most of the cases. But these settings must be applied on both sides, ESXi server and switch. Otherwise you have collisions and collisions (with a switch) should never occur if everything is working well.

Can you check the switch port if there are collisions occur?

AWo

VCP 3 & 4

\[:o]===\[o:]

=Would you like to have this posting as a ringtone on your cell phone?=

=Send "Posting" to 911 for only $999999,99!=

[mihalina]

hello,

thank you for replay. We have also checked switch is causing problems,

but it seems like that on switch everything is fine. We have now auto

negotiate on switch and on esxi (for local network). I also try to change the port on switch but it does not help.

I was having full duplex 100 mbit/s on switch and on nic for local network (vsphere).

Thanks!

[AWo]

but it seems like that on switch everything is fine.

Seems? Have you checked if there are collisions when you use autonegotiation?

AWo

VCP 3 & 4

\[:o]===\[o:]

=Would you like to have this posting as a ringtone on your cell phone?=

=Send "Posting" to 911 for only $999999,99!=

[mihalina]

Yes we checked on there are no problems (no collisions).

[mihalina]

Yes we checked and there are no collisions on autonegotitation.

[rickardnobel]

Can you do a "show interface fa0/XX" on the switch and post the result?

[mihalina]

Hello,

Here is a copy of show interfaces 0/14

FastEthernet0/14 is up, line protocol is up (connected)

Hardware is Fast Ethernet, address is 0007.84fd.32ce (bia 0007.84fd.32ce)

MTU 1500 bytes, BW 100000 Kbit, DLY 100 usec,

reliability 255/255, txload 1/255, rxload 1/255

Encapsulation ARPA, loopback not set

Keepalive set (10 sec)

Full-duplex, 100Mb/s, media type is 100BaseTX

input flow-control is unsupported output flow-control is unsupported

ARP type: ARPA, ARP Timeout 04:00:00

Last input never, output 00:00:01, output hang never

Last clearing of "show interface" counters 1d00h

Input queue: 0/75/0/0 (size/max/drops/flushes); Total output drops: 0

Queueing strategy: fifo

Output queue: 0/40 (size/max)

5 minute input rate 0 bits/sec, 0 packets/sec

5 minute output rate 1000 bits/sec, 2 packets/sec

4394391 packets input, 321765876 bytes, 0 no buffer

Received 3807 broadcasts (0 multicast)

0 runts, 0 giants, 0 throttles

0 input errors, 0 CRC, 0 frame, 0 overrun, 0 ignored

0 watchdog, 2202 multicast, 0 pause input

0 input packets with dribble condition detected

7569349 packets output, 2469807122 bytes, 0 underruns

0 output errors, 0 collisions, 0 interface resets

0 babbles, 0 late collision, 0 deferred

0 lost carrier, 0 no carrier, 0 PAUSE output

0 output buffer failures, 0 output buffers swapped out

[rickardnobel]

Here is a copy of show interfaces 0/14

So everything looks good here. There should be no duplex problem. The thing to look for on a duplex-mismatch problem is "CRC errors" on the full side and "late collisions" on the half-duplex side, there is none.

There is also no drops, either input or output.

So I do not think your problem is with the physical switchport, perhaps it is a logical problem. Could you do a printscreen of the network configuration of your host, that is the vSwitches and their connection to VMNICs?

[mihalina]

Because I new to ESXi could you explain how to see configuration of vSwitches and their connection to VMNICs?

Thank you!!!

[mihalina]

Did you mean this?

1. esxcfg-nics -l

Name PCI Driver Link Speed Duplex MAC Address MTU Description

vmnic0 0000:04:00.00 e1000e Up 1000Mbps Full 00:30:48:61:f7:34 1500 Intel Corporation 80003ES2LAN Gigabit Ethernet Controller

vmnic1 0000:04:00.01 e1000e Up 100Mbps Full 00:30:48:61:f7:35 1500 Intel Corporation 80003ES2LAN Gigabit Ethernet Controller

1. esxcfg-vswitch -l

Switch Name Num Ports Used Ports Configured Ports MTU Uplinks

vSwitch0 128 5 128 1500 vmnic0

PortGroup Name VLAN ID Used Ports Uplinks

Management Network 0 1 vmnic0

VM Network 0 2 vmnic0

Switch Name Num Ports Used Ports Configured Ports MTU Uplinks

vSwitch1 128 5 128 1500 vmnic1

PortGroup Name VLAN ID Used Ports Uplinks

172.31.1.1 0 3 vmnic1

?

Thanks!

[rickardnobel]

Can you also run:

esxcfg-vmknic -l

[mihalina]

~

1. # esxcfg-vmknic -l

Interface Port Group/DVPort IP Family IP Address Netmask Broadcast MAC Address MTU TSO MSS Enabled Type

vmk0 Management Network IPv4 212.13.248.130 255.255.255.248 212.13.248.135 00:30:48:61:f7:34 1500 65535 true STATIC

[mihalina]

And for explenation:

172.31.1.0 network 255.255.255.0, private network (problems)

212.13.248.128 network 255.255.255.248, pubic ip, everythink is ok.

212.13.248.130 is management ip. I was having management on private ip address, but I was lossing connection all the time.

[rickardnobel]

Are all your virtual machines on the Portgroup called "VM Network"?

Are there any VM:s on the portgroup that is (somewhat strangely) called "172.31.1.1" on the vSwitch1?

Are your problems at the moment connections with the virtual machines or with management of the ESXi host?

Are both your physical network cards (vmnic0 and vmnic1) connected to the same physical switch?

[mihalina]

NO, on portgroup VM Network is using public ips.

And the second portgrou name is 172.31.1.1. This VMs are usin local ips (172.31.1.xxx)

On first one are 4 vm machines (pubic ip) and on the second one are 3 vm machines (local ip) .

Yes the portgroup is called 172.31.1.1 and yes it is on vswitch1. I will name it differently :smileyhappy:

Ok, my VM which are having prive ip (on vswitch 1, and portgropu 172.31.1.1) and all are having connection. The problem is that tha packages are dropping. When I ping them 47/50 pings come throught. So that is why I have changed management ip on public ip. I was having problems all the time with vspehere connection, due to package loss.

No, nics are on different switch.

I was having Xen server on this machine installed and local network was not causing any problems.

[rickardnobel]

When you try to ping them, from which IP address do you do this?

Is it from a host inside the 172.31.1.x network or from any other? That is, do the traffic has to be routed?

[mihalina]

Hello ricnob,

I am pinging from switch on ip 172.31.1.101. I have also try ping from VM and pings are also losing.

I do not know for certainly if the traffic is routed but I think that it is. I will ask cisco man:)

[rickardnobel]

If you are ping from a 172.31.1.x address then there will not be any routing, just layer 2. So any routing problems should also be ruled out.

Next thing to check is the ESXi host CPU. Could the host be too overloaded?

Can you check through the vSphere Client and on the Performance tab - check the CPU load and then go onto "advanced" - "network" - "chart options" and deselect all counters and then select inbound and outbound dropped frames.

[mihalina]

I have acttached all files in rar.

Thanks!

[rickardnobel]

I see nothing strange on those performance charts either. No dropped packets and low CPU usage.

Do you get this dropped pings against all of your VMs on vSwitch1?

[mihalina]

Hello @ricnob,

yes I get dropped pings against all of my virtual host on privat network.

I am thinking if something is wrong with my NIC, but I had xenserver installed and every think was fine.

I will try to delete my virtual card and configure everything agin. We will see what will happen.

Do you maybe have any other suggestion where should I look?

Thanks for your help!

Miha

[rickardnobel]

Are both the VMNIC and the switch port set to auto now?

Have you verified if there is any error logs inside the VMs that could be related to the network drops?

Can you also check that there is not some CPU problem inside the guests, as if they are running on 100% cpu which could cause this.

If these things are ok, then I can not think of much more to check at this moment. It could be hardware related, or perhaps some issue with the vmware driver for the VMNIC brand.

[mihalina]

This is interesting thing. On switch is configured on auto mode. But if I put on auto mode in vsphere the connection auto set on half duplex.

[rickardnobel]

On switch is configured on auto mode. But if I put on auto mode in vsphere the connection auto set on half duplex.

It becomes "half" on the switch side if setting vSphere to "auto"? That is strange of course. It seems like the auto-negotiating is not working. If one side has auto and the other side does not respond (perhaps configured to not-auto), then the standard says that it shall become HALF.

So if the switch with auto becomes half duplex, then either your vSphere card is already set to static half/full, but not auto. Or the negotiating is for some reason not compatible. This was a problem many years ago, but these days most NICs and switches does this without problems.

[mihalina]

It is set on auto on bouth sides.

It also noticed that others are also having this problem

with same nic card (motherbord supermicro x7db3).

I will try to add different network card.

On Mon, 27 Sep 2010 07:36:53 -0700

[mihalina]

Here is a copy of my events:

Network connectivity restored on virtual switch

"vSwitch1", portgroups: "VM Network 2". Physical

NIC vmnic1 is up.

info

9/27/2010 12:01:33 PM

vmware.vmware

I will also try with another nic.

Thanks!

[rickardnobel]

I will also try with another nic.

It will be interesting to see what your results will be with a new NIC. Everything looks that it should work, so perhaps a hardware or driver issue is most likely.

[mihalina]

~

1. # esxcfg-vmknic -l

Interface Port Group/DVPort IP Family IP Address Netmask Broadcast MAC Address MTU TSO MSS Enabled Type

vmk0 Management Network IPv4 212.13.248.130 255.255.255.248 212.13.248.135 00:30:48:61:f7:34 1500 65535 true STATIC