The Team Expands

Hi All, my name is James Henderson. I have just joined the team here at myVMX and will be contributing alongside Kin Yung and Nuno Mota in providing articles, reviews, tips and tricks for a variety of different subjects. 

Some of the areas I work around include: System Center Suite (with a particular focus on Operations Manager (SCOM), Apple iOS, Mobile Device Management, Certificate Services and a variety of others. 

I hope my posts are of interest and I can't wait to get started.

Veeam Backup and Replication Cloud Enabled Edition

Veeam has just recently released backup and replication cloud edition. This allows you to use public cloud storage to store your backup as Veeam has written the required API to connect to these providers. Currently it supports the following 15 cloud storage providers listed below and I believe it will grow as the up take of the product happens:

Amazon S3
Amazon Glacier
Azure
Openstack
Rackspace
Google
HP Cloud
Clodo
Haylix
Scality
Dunkel
Tiscali
HostEurope
DreamObjects
GreenQloud
Seeweb
Connectria
Walrus
Mezeo


Veeam is a great backup product and I have been using it in my test lab to help me backup my lab machines. There are some questions which should be thought through before using a cloud storage provider for backup;

• Does your data have to be held within a country? region?  due to sensitivity? or restricts? As you may not know how these providers distribute or store your data you need to understand this in case it is breaking some laws for the data you have

• Most providers charge you very low cost to store the data but when you recover the data they charge you to transfer out. So check how often you need to restore and how much data you normally recover to give you a rough idea how much it could cost you over a period of time

• Do these provider offer alternative way to retrieve the data? ie sending you the drive?. Reason is that if you lost 5 virtual machines and you needed to retrieve 300GB data across the wire how long would that take? could you do it within your SLA you have with your customers/users?. If your primary data centre failed would you be retrieve the data somewhere else?. You don't want to be thinking about these issues when a disaster strikes

•  Are there support numbers you can call and speak to someone? or is it just via email?

• Can you just retrieve the data yourself or do you need to log a call first?, some providers archive to tape so a support ticket it required to be log so that they can load the tape for you, this means the SLA times will change as well for you and customer

• Does you current internet pipe have the capacity to take the backup transfer load as well?, do you need to upgrade? you don't want your backup and recovery disrupting your main business

• How secure is your data held at their data centre?, do system administrators at the provider have access to your data without your knowledge? are you in control of your data?

• Have the investment you made for your current backup infrastructure paid for itself now? was a lot of capital investment made for it and to change to the new cloud backup would mean you would need to write off the investment?

• What happens if the company goes bankrupt or shuts down?, can you still get your data?

These are some of my thoughts that you should consider before moving your backup to the cloud. I agree it could save you money i.e. no capital spend (some companies like that) but do check that the cost is not hidden in something else like needing to upgrade your internet pipe or the cost of retrieving data. Also it sounds good that you can move from provider to provider but in reality would you do it? given the amount of data you would need to move and more management overhead required to look after these providers. By all means test it with Veeam before you commit to changing your whole backup strategy to using cloud storage as this is a big jump in my opinion.

Here is the link to the new product from Veeam :
http://www.veeam.com/cloud-backup-vmware-hyper-v.html?ad=btn

YouTube Channel vmware technical publications

For people who would prefer to listen and watch there is a youtube channel for vmware technical publications

http://www.youtube.com/user/VMwareTechPubs

Released: Exchange Server 2010 SP3

Finally the much anticipated SP3 for Exchange 2010 has just been released! It brings the features and capabilities:

• Support for installing and deploying Exchange 2010 on Windows Server 2012;
• Support for Internet Explorer 10 to connect to Exchange 2010;
• Fixes;
• Coexistence with Exchange 2013.

Important: Exchange 2010 SP3 allows Exchange 2010 servers to coexist with Exchange 2013 CU1, which is scheduled to be released in February 27th.

 

Please also note that this Service Pack requires schema updates.

 

DOWNLOAD it here.

vsphere 5 netapp NFS disconnect bug

I have just read on two blog post by Cormac Hogan and Vaughn Stewart where there seems to be an issue with vSphere 5.x connected to Netapp using NFS. On occasions where there is high load (I/O) the datastores would disconnect or disappear for a period of time and then reappear. There is some work around available which are
1) Reduce the max queue depth down to 64
2) If you have license to use storage I/O control then this would help prevent the issue
3) Upgrade ONTAP to either 7.3.7 P2, 8.0.5 or 8.1.3

We are not currently experiencing these issues but we will be implementing option 1 in case it happens suddenly. In the long run we are planning the upgrade of ONTAP so that we can fix the issues.

Here are the blog posts from Cormac Hogan, Vaughn Stewart and VMware KB article

http://cormachogan.com/2013/02/08/heads-up-netapp-nfs-disconnects/
http://virtualstorageguy.com/2013/02/08/heads-up-avoiding-vmware-vsphere-esxi-5-nfs-disconnect-issues/
http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=2016122

Disk timeout value for guest OS on Netapp NFS storage

There is a setting which should be applied to all VM guest OS if you are using Netapp NFS storage as your datastore for your VMs. This setting increases the disk timeout value to 60 or 190. The setting is to prevent VM’s from blue screen, restart or lose connectivity if a netapp cluster fail over does occur. Although failovers usually take less than a minute this is a precaution so that the VMs stay up and delay the write to disk for up to 190 seconds. You can get the settings by installing VSC which will give you the ISO to mount onto the vm to apply the change.

Here is the reg value for Windows OS as that is what we primarily use:
HKLM\SYSTEM\CurrentControlSet\Services\Disk\TimeoutValue to DWord 190

Vmware Fling Inventory Snapshot

Recently we have been rebuilding our vCenter to become a virtual machine. Rebuilding vCenter is easy but to take over all the roles, permissions and VM folders it could be a real pain. I found this little app on vmware flings called “inventorysnapshot”.

This application will capture all details including roles, permissions and VM folders. It will then create a powershell script which you then run against the new vCenter. The script would populate all the folders, add the host from the old vCenter to the new vCenter. Move all the VM’s to the correct folder and assign all the permissions!!! How cool is that.

Be aware though, use it when you are doing a complete migration as it removes the host from the old vCenter and add to the new one so notify people before you do the move ^_^. Turn off DRS and HA from the old cluster as well in case

The software can be obtained from here: 

http://labs.vmware.com/flings/inventorysnapshot

Exchange 2010 dropping network packets( vmxnet3)

At my work place we have our Exchange 2010 running completely on VMware. We use SCOM as our main monitoring tool so we purchased Veeam nWorks to pass on all the metrics to SCOM and generate alerts where needed.

Over the pass few weeks we started to receive alerts regarding dropped packets from an host. At first we thought we had over subscribed the bandwidth but after looking at performance charts within vCenter for the host. We noticed that we had been having drop packets for a while but was unaware of the issues as it never hit the threshold for SCOM to send an alert for us. After seeing this we quickly checked across all our host in case we had an underlying issue waiting to happen and found out is was only happening on host which was hosting exchange. Further investigation we discovered it was coming from just our mailbox servers.

There was an article released from VMware regarding this problem. The issue is due to the fact we are using vmxnet3 network cards and when Exchange has a burst of traffic the network card buffers can't take it. The workaround or fix is to increase buffer on the network card. The process is described in the KB article:

http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=2039495

New Exchange License Cmdlets

We all know the old script to count the Client Access Licenses [CALs] in Exchange 2010 was never 100% accurate... With Exchange 2013, two new cmdlets were introduced: Get-ExchangeServerAccessLicense and Get-ExchangeServerAccessLicenseUser.

Get-ExchangeServerAccessLicense

This cmdlet returns a list of licenses in use in an Exchange 2013 organization, with the legal name of the license as defined in the Microsoft Product List.

If, for example, you want to check all the licenses in your organization, simply run the following cmdlet:

Get-ExchangeServerAccessLicense

The possible license names are:

• Exchange 15 Standard CAL
• Exchange 15 Enterprise CAL
• Exchange 15 Server Standard Edition
• Exchange 15 Enterprise Edition

Get-ExchangeServerAccessLicenseUser

This cmdlet, on the other hand, returns a list of unique users (their fully qualified domain name [FQDN] or primary SMTP address) representing an estimate of the licenses required when it is run.

To check how many users need a Standard CAL, for example, you run the following cmdlet:

Get-ExchangeServerAccessLicenseUser -LicenseName “Exchange Server 15 Standard CAL”