XenApp/PVS Global Farm Overview

Since there was a lot of interest in the last Visio I posted, I thought I’d share another.

These diagrams outline high level overviews of a global XenApp w/PVS deployment, with XenApp zones and PVS sites in each datacenter. Each XenApp zone has two data collectors/XML brokers, PVS-Streamed OU-based worker groups. Each geographic region has a corresponding StoreFront Store (directed by host name):



The PVS farm configuration is very similar, consisting of sites in each datacenter to stream XenApp workers for each XenApp zone in that datacenter, with the master database homed in the US datacenter:



The intent of these overviews are mainly to demonstrate how the XenApp and PVS farms interact in a global zone/site architecture. I’ll share some overview diagrams of XenApp zones and PVS sites in another post. Hope you enjoy!


XenDesktop Session Launch Hypervisor Interactions

I got an email recently asking if I knew whether or not a XenDesktop site takes a hosting unit’s load or availability into consideration when brokering session launch requests, especially reconnects to desktops that were ‘In Use’ when a host goes down. This question was posed in the context of Desktop Groups with catalogs that are spread across multiple hosting units.

The simple answer to this question is no. XenDesktop’s interactions with the hypervisor (via the Hypervisor Abstaction Layer) were always intended to be used for power action/status, and MCS/PVS related cloning activities. When a XenDesktop site selects a ‘worker’ to fulfill a session launch request, it only looks at the worker’s registration status, and not that of the host that the worker guest VM is running on.

That said, the selection process for the next available worker is determined via stored procedures. To find out what the ‘next available’ worker is going to be in a XenDesktop 5.x site, you can run following T-SQL against the database, specifying the Desktop Group UID in the first line:

declare @DesktopGroupUid int = 1
declare @Readiness int = 3
declare @Uid int

 update Top(1) chb_State.Workers
 set @Uid = W.Uid
from chb_State.Workers W
            with (readpast,
         where DesktopGroupUid = @DesktopGroupUid
           and LaunchReadiness >= @Readiness
           and SinBinReleaseTime is null;select * from chb_State.WorkerNames
where Uid = @Uid

This script will return the name of the machine that the site will use to satisfy the next pooled-random session launch request to the specified desktop group, and doesn’t care what’s going on with the hosting unit where the worker lives. The site is only concerned with the worker’s registration state, and could care less if the power state of the VM is On, Off, or Unknown, much less does it care about the load of the hosting unit where that machine is running.

To that point, if a worker continues to register when a hosting unit connection becomes inaccessible (vCenter is down, but not the ESX host, for example), the desktop will still be available for session launch, but not for power management. This scenario can cause problems, such as ‘tainted’ workers that don’t get powered off after use, and end up in the unfortunate sounding ‘SinBin’. This process is only temporary, and is only corrected after the machine is rebooted by the XenDesktop site (check out the CTX article I wrote for more info).

As far as a scenario where a session was ‘In Use’ when the host goes down, the broker reaper site service will eventually clean up the failed worker when the ‘DDC Ping’ times out (controlled by the ‘HeartbeatPeriodMs’ value on the DDC running the reaper service). So, by default, you could potentially get into a situation where reconnects for ‘In Use’ session will keep selecting the failed worker until the reaper cleans it up. While this shouldn’t take longer than 5 minutes with the default heartbeat value, it may cause problems if there are frequent outages or service interruptions between geographically dispersed datacenters.

To work around the ~5 minute functionality gap of hosting unit availability awareness, as it relates to session launch anyways, one could easily trigger a XenDesktop PoSH script in the event of an outage (and the reverse when the outage is recovered) to toggle the ‘maintenance mode’ flag on any workers on a failed host. I’d like to hope that the XenDesktop product team has at least considered the potential for expanding the site’s visibility into the status of a guest VM’s host, and would love to see ‘smarter’ brokering logic such as thing in future releases.

XenDesktop 7 – First Thoughts

Citrix hosted an amazing event last week, and outlined a distinct roadmap of their 2013 strategy. They placed a strong emphasis on mobility with some updates to their Zenprise acquisition (XenMobile, aka Worx), and announced the first implementation of Project Avalon in the form of ‘XenDesktop 7’. Since I’ve spent a lot of time with XenDesktop (both IMA and Storm based) and XenApp, I thought I’d share my general impression of XenDesktop 7 as it relates to achieving the goals set forth by Avalon.

First off, the unification of XenDesktop and XenApp was a necessary evil based on Citrix’s decision to combine the management and provisioning of  ‘desktops’ & ‘servers’ (SBC and VDI) within the same console. Through what Citrix is calling the ‘FlexCast Management Architecture’ (Storm+RDS), they are replacing ‘IMA’, which was used for all versions of XenApp, as well as XenDesktop versions prior to Rhone (Barossa, Sonoma, Rioja, Bordeaux, Medoc, etc.).

This change is a great move in terms of farm design, scalability, and stability. In my opinion, the Storm framework is easier to install, troubleshoot, and support than IMA (written in .NET, readable database, excellent SDK, better logging, etc), and should be familiar to anyone who has worked with XenDesktop 5.x. The site is just as dependent on availability of the central database as in XD5 (no local host cache), which means no zones, data collectors, or any other sort of ‘master’ server (the database is the master). All of the same ICA/HDX functionality is still there (plus any new additions), as is the policy engine and brokering functionality.

I’m not too fond of the licensing model which provides published Windows client OS in the least expensive edition, whereas Windows server OS requires a more expensive license. I suppose that’s representative of Citrix choosing to call Excalibur XenDesktop instead of XenApp, though I never really thought of this distinction since I assumed it was called XenDesktop because they used the Storm site architecture (now called FMA). I’m also concerned about feature parity with XenApp, and am sure there will be more than a few features that either don’t live up to XenApp, or just aren’t there yet.

At the end of the day I’m excited about XenDesktop 7, as it provides an easier product to sell. There’s no more worrying about whether or not you need to publish apps from Windows client or server OS (besides the licensing), and all of the management and provisioning (except for Provisioning Services :)) is done in a central console. The new Director looks fantastic, and the refreshed Studio is much more responsive and elegant than that of XenDesktop 5. Also, my SiteDiag tool (Site Checker v2.0) was designed to run on the Excalibur tech preview, and I’ll be sure to get it working for XenDesktop 7 once its released.

I get the feeling that the rest of the Citrix community is generally as excited about XenDesktop 7 as I am, but I guess we’ll see how it plays out once we start implementing it!