Jump to content

Recommended Posts

Posted

Morning all,

 

We have the following setup:

 

1 x VMware host running a DC which also acts as the fileserver for network shares (homefolders, resources, etc)

1 x VMware host running another DC plus several other servers (AV, Exam software etc)

1 x physical DC. (plus several other physical servers)

 

All are connected via 1Gbps links to 2 x Cisco 3750 stacked switches. The connections from the VMware hosts through to the storage units (1 x SX80 2TB unit for virtual machines and 1 x DX60 12TB unit for shared folders) are all VLanned off.

 

All runs well apart from a couple of random times throughout the day whereby the shares become very, very slow to access for around 10 -15 minutes, these can be at any time of the day and there seems to be no fixed pattern.

 

The drives in the storage units are showing no issues whatsoever, On the fileserver itself shares load no problem, but from any other machine access is very slow. RAM, CPU and network usage are all very low on all of the servers, the switches are showing no spikes in usage (the same applies to the VMware hosts themselves). It isn't down to AV as I've tried turning this off. the network isn't being flooded and there is nothing showing as starting up on the server in eventvwr. I am totally flummoxed.

 

Has anybody any idea what else I can try/test?

 

Regards

 

nick

Posted

Snapshots? Or vMotion? vmware fully up to date?

 

What if you completely uninstalled the AV on the server (just disabled, the filter drivers are still in the data path and active to some degree)?

 

Also try these:

List of currently available hotfixes for the File Services technologies in Windows Server 2008 and in Windows Server 2008 R2

AppLocker path condition does not work when a file name contains international characters in Windows 7 or in Windows Server 2008 R2

(assuming you are up to date with your Windows Updates, this will move you to the latest kernel build, which has significant caching improvements and changes to support other bug-fixed components higher up the stack)

Slow data transfer speed in Windows 7 or in Windows Server 2008 R2

And then

http://support.microsoft.com/kb/2750841

(which contains further improvements to the TCPIP Stack)

 

Push the lot out to all your clients and servers. (scripting the install and removal with powershell is a copy/paste job thanks to Google)

  • 1 month later...
Posted

Thanks for the reply Psydii, I've now managed to get it all up and running as it should!

 

I had VMware and MS looking into the issue. VMware could see no problems with the setup, I was convinced that there was an issue with TCP/IP but MS were not in agreement, the 2 departments I was in touch with tried a few hotfixes, but nothing worked (and in fact the iSCSI hotfix seemed to make it worse). As time was ticking on we moved the shares to a fresh server and connected the SAN as a raw drive as opposed to iSCSI and access to the shares had no problems all was good again, HOWEVER, the server to which they were connected was a DC and suddenly log on times took up to twenty minutes!. On the DC I reset TCP/IP and winsock turned off autotuning and applied one of the hotfixes above (although it applied to IPV6, nothing ventured nothing gained) rebooted and VOILA, Bobs your aunty Jane! all seems well again!

 

Suffice to say I shall be rebuilding the DC when we get a chance!

 

Thanks

 

manick

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now



×
×
  • Create New...