Average size of NFS TCP packets received from host: esx203-NFS is 224....

[__ple_16414]

Hello,

We have a customer with VMware vsphere 4 servers and Datastore NFS on NetApp 3140 MetroCluster.

The following warning come back several times in the log and when we excecute a netdiag -ve.

What can we do or should we not take those warnings into account ? :

"Average size of NFS TCP packets received from host: esx203-NFS is 224.

This is less than the MTU (1500 bytes) of the interface involved in the data transfer.

The maximum segment size being used by TCP for this host is: 1460.

Low average size of packets received by this system might be because of a misconfigured client system, or a poorly written client application.

Error 406

Press enter to continue"

Thanks a lot for your anwsers

Philippe

[danielpr]

I am also experiencing the same issue for my two different VMkernel. But the next time I do the diag the error goes off.

Performing transport layer diagnostics.....
Checking for problems with current TCP connections...

Average size of NFS TCP packets received from host: xx.xx.xx.xx is 275.
This is less than the MTU (1500 bytes) of the interface involved in
the data transfer.
The maximum segment size being used by TCP for this host is: 1460.
Low average size of packets received by this system might be
because of a misconfigured client system, or a poorly written
client application.
Error 406
Press enter to continue

Average size of NFS TCP packets received from host: xx.xx.xx.xx is 578.
This is less than the MTU (1500 bytes) of the interface involved in
the data transfer.
The maximum segment size being used by TCP for this host is: 1460.
Low average size of packets received by this system might be
because of a misconfigured client system, or a poorly written
client application.
Error 406
Press enter to continue

Checking for problems with recent TCP connections...
UDP is OK

Thanks

Daniel

[m_lubinski]

yes, I also see these messages (similiar)

Average size of NFS TCP packets received from host: x.x.x.x is 7896.
This is less than the MTU (9000 bytes) of the interface involved in
the data transfer.

as well as something like this:

The interface nfs-vif1-2 seems to be receiving an excessive number of packets
for protocols not supported this system.
If this problem persists, please contact your network administrator.
Error 209
Press enter to continue

[__ple_16414]

Thank you both.

Daniel, I must do something or consider it as a warning not important ?

Thanks for your answer,

Philippe

[danielpr]

Phil,

I am more worried about the initial packet size of MTU. We should make sure netdiag is always OK for system to function properly.. When I looked at the Error number it says poorly configured client system and also something to do DNS settings.

I don’t have any idea how to get rid of this MTU error. We will wait for experts to help us here.

Thanks

Daniel

[__ple_16414]

Hello Daniel,

You say "We will wait for experts to help us here."

Do you know some NetApp experts that can help us ?

Now, we are many to have this "issue"

Philippe

[danielpr]

Hi Phil,

I think this is due to tcp windowing.. The intial MRU and MTU negotiation goes lower  MRU and eventually getting increased or reaches to MRU 9000.However this TCP datagram size should be 9000 in normal cases, instead  its goes to 1xxx or lesser than it. Which will not have much impact because once again MRU goes up to 9000 and netdiag doesn’t show any error.

I am tracking this with Unix/Lab admin in my test Lab. If they find any issues/solutions I will let you know .

Thanks

Daniel

[__ple_15582]

Hi Daniel, Hi m.lubinski

Thks both for your help.

Philippe

[m_lubinski]

Hi,

I got today answer from Netapp support about these messages (402,406)

You called to get assistance with this issue. I had him add the log to the case and I looked through this. All of the errors are 402 and 406 and as all of the messages use this verbiage as apposed to message or notification this particular output simply decided to use error for the messages. These errors simply mean that the packets expected to be received by the are expected to be 9000bytes in size and they at times are smaller that that. This is nothing to be concerned with as it simply means that if the applications do not have enough information to move 9000 bytes at that point and as such sent a smaller packet. I did speak to a senior engineer and he agrees with this assessment.

[vmsjaak13]

I'm also seeing the message "Average size of NFS TCP packets received from host: 10.1.10.2 is 224" quite a lot.

Does anyone have additional information ?