It is a small thing but a little consistency would be nice… I have one example below, just one, I am being nice.
DavesTechnology: Having worked in IT for many years I work with lots of customers and different technologies. Day to day it should be easy but there is always strange stuff that happens with both. Not often it just works. Let me tell you why... Dave's Technology DavesTechnology
It is a small thing but a little consistency would be nice… I have one example below, just one, I am being nice.
Error MFCOM Service error when running doing the initial Citrix discovery (w28kr2/XA6.5)
“Errors occurred when using <server name> in the discovery process.
An unexpected error occurred. Check that the server name is correct, that the server is on, that Citrix Presentation Server is installed on this server, and that the Citrix MFCOM Service is running.”
There are multiple possible causes for this issue. The most common are as follows:
See below, it will show this error as the IMA service does not start because it is not configured… Run the XenApp Role Manager and follow the bouncing ball.
Still a problem?
Users group on the server you are connecting to. The recommendation is to do this using groups.
Another option to check.
On the server, set the DCOM Default Impersonation Level to Impersonate by following these steps:
a. Go to Administrative Tools, Component Services, Computers, Right-click My Computer and select Properties.
b. Select the Default Properties tab.
c. From the Default Impersonation Level drop-down list, select Impersonate and click OK.
Reboot and try again.
http://support.citrix.com/article/CTX112853
Update:1 – corrected some numbers. Thanks @jasonboche
Let me just say I love VMware, particularly ESX and have worked with it from before GSX was in shorts, back in the workstation days, when only para-virtualisation existed. I have rolled out ESX 2/3/4 farms (no 5 yet)… I have never had a purple screen of death, I have never had to rollback a workload to hardware, I have VMed Exchange 2k/2k3/2k7, DCs 2k-2k8r2, file & print, SQL servers, Citrix servers they all run great on ESX/ESXi.
But do you need it? Or are you after a solution has all the features, are you after the Rolls Royce? What are you really trying to do? Are you exotic or somehow special?
You want ESXi that's fine, go ahead I don't get paid either way, but then nor does anyone so please do.
Lets just think about what server virtualisation does (as of todays date):
Feature | ESX/ESXi | Hyper-V | XENserver |
Bare-metal architecture | Yes | No, but core | Yes |
VMotion like | Yes | Yes | Yes |
Small footprint | Yes | No, but core | Yes-ish |
Cluster (pool) | 32 nodes | 16 nodes | 16 nodes |
CPU virtualisation | Yes | Yes | Yes |
RAM Support Host | 2TB | 1TB | 1TB |
RAM support VM | 1TB | 64GB | 128GB |
RAM overcommit | Yes | Yes | Yes |
NIC teaming | Yes | No*/vendors | Yes |
VM RAM Page sharing | Yes | No | No |
Ballooning | Yes | Yes | Yes |
Capacity prioritisation | Yes | Yes-ish | Yes-ish |
Traffic Shaping | Yes | No | No |
Virtual NIC | Yes | Yes | Yes |
Virtual switches | Yes | Yes | Yes |
VLAN tagging | Yes | Yes | Yes |
Dynamic volume resizing | Yes | Yes | Yes |
Raw device mapping | Yes | Yes | Yes |
LUN management | Yes | Yes-via vendors | Yes-add on |
Guest Windows | Yes | Yes | Yes |
Guest Linux | Yes | Yes-limited | Yes-good |
Guest Other | Great | OK | Good |
Paravirtualisation | None (good!) | Yes, LAN/Disk | yes-ish |
Distributed Power Mgnt | Yes | Some | Some |
Wake-on LAN | Yes | No | No |
There are hundreds more features such as “Boot from SAN” which are specific or particular so I have not listed them above, if YOU need them they are critical but lets just focus back on the 99%
So to look at this list above there are some clear areas where ESXi wins out, specifically on the very large scale, telco scale, intensive power saving, dare I say cloud providers… But if you are one of the people who just need a bunch of VMs per server for general workloads any of the three products above is going to work fine. Scale up the CPU, RAM, Network cards and you can go to higher density?
The limits are disappearing.
So now lets look back at yourself, do you really need ESX? Maybe you could save some money… Look around.
Just before signing off, read this: http://www.thincomputing.net/2011/03/07/how-many-users-can-i-host-per-server-with-remotefx-for-hyper-v-and-what-is-the-cost-per-user/
This document is the source for the core of this post:
“vmware-vsphere-features-comparison-ch-en.pdf”
Intel: “xeon_7500_Virtualization_solbrief.pdf”
* Microsoft say dont use teaming, but vendors support it. Buyer beware… Microsoft Support Policy for NIC Teaming with Hyper-V - http://support.microsoft.com/kb/968703
Without an intervention from you, users settings and user files are stored in the local user profile, under the Users folder on the local drive. “C:\USERS\etc”. I think we all know we cant trust users to backup so lets look at it.
There are two technologies to ‘fix’ this situation, Roaming Profiles and Folder Redirection.
Folder Redirection lets you redirect the path of a folder to a new location. The location can be another folder on the local computer like a D: or a directory on a file server. to the users it is as if the documents were based on a local drive. The documents in the folder are available to the user from any computer on the network and offline via offline folders.
Advantages of Folder Redirection
If users log on to different computers on the network, their data is available.
Offline Files (which is turned on by default) gives users access to the folder even when they are not connected to the network. This is for people who use laptops.
When it is stored in a network it can be backed up…. Nice idea huh?
When using Roaming User Profiles, you can use Folder Redirection to reduce the total size of your Roaming Profile and make the user logon and logoff more quicker for the user.
You can use GPO to set disk quotas, limiting how much space is taken up by user profile folders.
You to select the location of the redirected folder on a network or in the local user profile:
Folder Redirection in Windows 7 improves first-time Folder Redirection performance because the the computer redirected folder data in the background, not just at logon. However the first time a user logs on, Offline Files moves all files and folders from the server to the local cache. The user is blocked from logging on to the computer during this task. Then, Offline Files synchronises from the local cache with the redirected user folder on the server. . So turn this on at the initial deployment, not later and be careful of WAN links.
Offline Files synchronises new and changed files and folders from the local computer to the server when the network becomes available or in the background when the connection is slow.
Slow link detection works via Network Location Awareness (NLA). This networking layer service allows applications, like GPOs, to request networking information from the network adapters in a computer, rather than implementing their own. NLA monitors the existing traffic of a specific network interface. This provided two important benefits:
A shared folder automatically transitions to the slow-link mode if the round-trip latency of the network is greater than 80 milliseconds, or as configured by this policy.
After transitioning a folder to the slow-link mode, Offline Files synchronises the user's files in the background at regular intervals, or as configured by the 'Configure Background Sync' policy. While in slow-link mode, Windows periodically (every 2 Minutes) checks the connection to the folder and brings the folder back online if network speeds improve.
It is not always plane sailing occasionally the laptop users loss where it was, don't let it sync or don't know where files are, but this is outwaited by the value of having their information backed up.
From the Start menu, type sync and it will bring up the sync centre to allow you to see what is going on.
This information is generally from here:
http://technet.microsoft.com/en-us/library/cc732275.aspx
http://blogs.technet.com/b/netro/archive/2010/09/15/slow-link-detection-for-offline-files-in-windows-vista-sp2-amp-windows-7.aspx
http://blogs.technet.com/b/askds/archive/2009/10/23/group-policy-slow-link-detection-using-windows-vista-and-later.aspx
If you run a bunch on VMs in VirtualBox and use NAT to keep consistent addresses changing the active network card (from wireless to wired etc) on the computer is a pain as you need to go into the VM and click disconnect network, ok, connect network ok.
Well that is way to may steps for my liking. Here is the command line to pop and icon in the desktop or run as a task/script.
vboxmanage controlvm "WORK SOE" setlinkstate1 offIf you want to change the network adaptor connection (ie, not connected or to the NAT), this is how.
vboxmanage controlvm "w2k8r2dc1" nic1 null
vboxmanage controlvm "w2k8r2dc1" nic1 nat
You can also start and sleep them on the command line.
vboxmanage controlvm "nameOmachine" savestate
vboxmanage startvm "nameOmachine"
And just to suspend the laptop from the command:
sudo pm-suspend
Most large organisations have two SANS and many larger organisation still have two or more data centres.
Lets just start with this is not a cloud, it is a data centre or separate SAN, OK that is off my chest. Bloody private cloud, this and that, don't get me started, “Why I oughta…”
You can utilise this architecture with VMware ESX/vSphere as active-active or active-passive, but I don't see value in active-passive so lets not speak of it again, it is not cheap, not effective and not quick. So back to Active-Active.
To build a single ESX cluster over 2 SANs (or 2 data centres) is basically the same, but as you reach out of the single data centre you need some extra configurations such as:
The value of this is protection from a SAN failure and all hardware used. If this is across data centres then you also have DR knocked on the head (your mileage may vary).
There are some issues such as a power outage in DC1 will cause all the VMs to restart in DC2 via HA so it is not true protection from massive outages. Secondly if you use vMotion to move the servers to the other nodes in DC2 the SAN writes are now subject to the latency of this link, so there can be a performance penalty. Additionally DRS may need to be manually updated to be aware of this configuration.
This is how it would look (minus some SAN switches etc.)
Here are the VMware prerequisites:
Here is Cisco discussing this:
Here is a good blog on this:
http://www.van-lieshout.com/2009/11/geographically-dispersed-cluster-design/
To clarify the distinction between upgrade and migrate:
XenApp 4/4.5 –> XenApp 5
XenApp 5 –> XenApp 6
To transition to XenApp 6, administrators must migrate to a new farm and have the following options:
Transferring settings from the legacy farm is best performed by installing the XenApp 6 Migration Tool on a XenApp 6 server and directly importing the settings by pointing at a server in the legacy farm. It is possible to import all settings or be selective. If the existing farm is based on server or application folders, these settings can be exported and designated so that they will be identified as Worker Groups within the new farm. Additional command-line entries are available to include, exclude, or modify specific settings. The administrator must have full administrative
rights in the target XenApp 6 farm and a minimum of view-only rights in the source legacy farm.
Prior to initiating the migration of settings, Worker Groups should be designated so that server and application silos can be assigned appropriately. This provides administrators with additional flexibility and enables a significant number of new servers to be added to the new XenApp 6 farm at once using a tool such as Citrix Provisioning services.
XenApp 6 –> XenApp 6.5
XenApp 6.5 implementations must be based on clean Windows Server 2008 R2 operating system installations as Citrix does not support operating system upgrades. Thus, there is no upgrade path to XenApp 6.5 from XenApp 5 or earlier products. In addition, a XenApp 6.5 farm can only encompass XenApp 6.5 servers - mixed farms are not supported.
I am of course an open source fan, but when it doesn't work, it is time for some proprietary solution. IE, Webex did not work with Icedtea Java Plugin and OpenJDK. Error "applet not initialized” at the bottom of the browser window.
So either apt-get remove or ubuntu software centre, remove OpenJDK and IcedTea java components.
then copy and paste the following:
cd
wget https://raw.github.com/flexiondotorg/oab-java6/master/oab-java6.sh -O oab-java6.sh
chmod +x oab-java6.sh
sudo ./oab-java6.sh
sudo apt-get upgrade
Open the browser and go to:
http://java.com/en/download/installed.jsp
It was ask for a plug in, select the IcedTea java plug and you should be sweet.
Furth (dated) details here: