We ran into this problem in a big way as we use NFS quite extensively.   In our case, it only seems to be applicable to our UCS based clusters.     We have yet to see the issue on one of our HP Proliant DL based clusters.   The bug is strange in that it surfaces more some UCS environments than other UCS environments.    Of course, a lot of variables to consider when comparing our UCS environments to one another but the common denomintor in our case is certainly UCS.
 
The good news is the datastore disconnect seemed to only occur for the ESX host responsible for generating the heavy load.   In other words, if your cluster consisted of 5 ESX hosts and a VM on host1 was generating heavy load, the APD/disconnect would only occur on host1.    The NFS datastore seemed to be happy on hosts 2-4.

We originally discovered the issues when Stroage vMotions would fail for no apparent reason.   Later, we found that we can generate the problem fairly easy from within a Linux VM.   Simply running a dd if=[file on vmdk hosted on nfs datastore] of=/dev/null bs=64k was enough to generate the issue in most cases.   At that point, we knew it wasnt specific to SVM.   Generating the issue via DD was also helpful because SVM creates a lot of noise in the vmkernel log so it was easier to see the problem occuring when we used DD.    
 
The problem seemed more likely to occur when the datastore was handling heavy read traffic.  Less likely to occur when handling heavy writes.    We troubleshot this for weeks and an L3 gentlemen from NetApp eventually recommended downgrading to 5.5 GA because they had been seeing some issues with U1 as well.    We dont see the issue with 5.5 GA.  We also dont see the issue with 5.0.
 
 


On Thu, Apr 24, 2014 at 4:15 AM, Boer, Arnout <Arnout.Boer@netapp.com> wrote:

See:

http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=2076392

-Arnout

 

 

From: toasters-bounces@teaparty.net [mailto:toasters-bounces@teaparty.net] On Behalf Of Marcus Nilsson
Sent: Saturday, April 19, 2014 7:07 AM
To: toasters@teaparty.net
Subject: RE: Vsphere 5.5u1 vs NFS

 

We also had a customer run into this.

 

BR Marcus

 

 

From: toasters-bounces@teaparty.net [mailto:toasters-bounces@teaparty.net] On Behalf Of Ed Morgan
Sent: den 18 april 2014 17:46
To: Blackmor, Chris; toasters@teaparty.net
Subject: Re: Vsphere 5.5u1 vs NFS

 

I¹ve seen it this on a new Flexpod I was rolling out, 5.5U1 on cDOT
8.2.1GAŠ Pretty much exact symptoms as the NFS Queue issue that was around
last yearŠ

Have reverted to vanilla 5.5 until VMware put something definitive out...

On 18/04/2014 15:34, "Blackmor, Chris" <chris.blackmor@exchange.amd.com>
wrote:

>I asked a friend at VMWare about this. She said that there is a bunch of
>buzz about it but to her knowledge there have only been 3 reported cases
>of this issue. Not sure if this is much ado about nothing (or almost
>nothing). Anyway, take it as third hand info, for whatever that is worth.
>Chris
>
>
>On 04/17/14 22:42, Mike Gossett wrote:
>> I am running it in production in 2 environments with no issues.
>>
>> Sent from my mobile phone
>>
>>> On Apr 17, 2014, at 10:11 PM, Mike Horwath<drechsau@gmail.com> wrote:
>>>
>>>
>>>> On Apr 17, 2014, at 7:45 PM, Andrew Laurence<atlauren@me.com> wrote:
>>>>
>>>> Word is going around that vSphere 5.5u1 causes lots of NFS
>>>>disconnects. Has anyone here seen this?
>>>> http://datacenterdude.com/vmware/nfs-disconnects-vmware-vsphere/
>>>
>>> My lab has been good so far.
>>> _______________________________________________
>>> Toasters mailing list
>>> Toasters@teaparty.net
>>> http://www.teaparty.net/mailman/listinfo/toasters
>>
>> _______________________________________________
>> Toasters mailing list
>> Toasters@teaparty.net
>> http://www.teaparty.net/mailman/listinfo/toasters
>_______________________________________________
>Toasters mailing list
>Toasters@teaparty.net
>http://www.teaparty.net/mailman/listinfo/toasters
>


Ed Morgan | Technical Consultant

 

T:
M: 07875 628794
E: ed.morgan@ansgroup.co.uk

 

ANS Group London Office
18 King William Street, London, EC4N 7BP
www.ans.co.uk

 

GBDataCentre



 

Disclaimer : The information contained in this communication from Ed Morgan sent on 2014-04-18 at 16:46:11 is confidential and may be legally privileged. It is intended solely for use by toasters@teaparty.net and others authorised to receive it. If you are not toasters@teaparty.net you are hereby notified that any disclosure, copying, distribution or taking action in reliance of the contents of this information is strictly prohibited and may be unlawful. ANS Group plc Terms & Conditions apply, All Prices are subject to VAT, Expenses excluded, E&OE

ANS Group Plc - Privacy Policy - Registered Office is Synergy House, Manchester Science Park, Manchester, M15 6SY. Reg No. 3176761. (Registered in England & Wales)

 


_______________________________________________
Toasters mailing list
Toasters@teaparty.net
http://www.teaparty.net/mailman/listinfo/toasters