- Published: Monday, 23 March 2015
On 6 March 2015 we organized our first DuCUG event for 2015. Already our 7th event. Again the event was well visited again and had again several great speakers. Also we had several people who spoke for the first time as such an event, one of our target as the DuCUG to get new speakers on stage. In this blog I wrote down my experiences of the day.
Our chairman Jon opened the event by mentioning the sponsors and announcing the agenda once again.
The first session was presented by Jits Langendijk about setting up Shared Service Center. Jits started with a quick overview of the organization, the underlying infrastructure and the product used (vSphere, XenDesktop 7.6, PVS 7.6, RES WM, App-V 5, RES AM, NetScaler, SafeNet Cloud 2FA and Fat Clients). Jits continued with the challenges they encountered during the project. One of the challenges was that the organization could not keep up. Also three organizations of the six original were going to cooperate which each other and a seven organization was added. Also the current infrastructure of the different organization were pretty different. Next topic was the technical realization (24 Delivery Group, phase approach per organization including phased migration of applications, data, back-end systems and the target reached at this moment). Jits continued with alternatives that could be used like: legacy environment via the central StoreFront, vDisks migrating to the central environment, client devices to 3 environments, VDA re-installation of the current XenDeskop environment. Jits shared his lessons learned about SQL mirroring, applications, NetScaler live migration, correctness of information and politics. The session ended with some answers and questions.
After the break we continued with the session of Ingmar Verheij with the title What’s new from Citrix Summit. Ingmar started with the vision of Citrix, with a nice referral to a video Citrix created in 2001 which actually became mostly reality nowadays. Ingmar started with the Software Defined workspace vision of Citrix (software layer to get the end user more productive, secure workspace delivery). Ingmar mentioned the acquisition of Octoblu (middleware between device and a service). Ingmar started mentioning the announcements: Receiver X1 (unified, dynamic, customize) and also showed in practice. Next the X1 Mouse was discussed by Ingmar (prototype mouse by Citrix for Receiver for iOS). After the nice way to give away a X1 Mouse Ingmar continued with the integration of Framehawk (Tech Preview Q2) and Lync 2013 Optimizations. Next topic was the stack in the data center: XenServer (6.5 release, 64bit, better performance), XenDesktop/App 7.6 Online Migration Tool (beta currently), departmental layering in an upcoming release without a filter driver / VHD chain (AppVolumes, Unidesk alternative), Session Recording (this quarter) and VDA Linux. Ingmar continued with the upcoming tech preview of DesktopPlayer for Windows and the XenMobile 10 release. Ingmar ended with the Workspace Cloud and WorkspacePod (preconfigured/preinstalled with two parnters dataON and HP) based on Sanbolic on Hyper-V.
Next session was Why we don’t need Citrix support (at all time) troubleshooting the FMA by Bas van Kaam. Bas started with the three steps in a troubleshooting process: investigate, analyze and implement. Bas continued with some tips out of the troubleshoot methodology: don’t try to guess the solution, don’t assume anything, isolate the problem, create a to read list, categorize/prioritize information, try solution with the latest impact first, write down what you have tried, share information and ask for help. Next topic were the tools available where Bas selected five tools for more information: Studio, Director, PowerShell, Citrix SCOUT and Insight Services. Bas touched quickly the FMA architecture before going further with Citrix Scout (exists of CDF-Control, XDPing, XDDBDiag, DC Logging, Events and Key data points). Bas continued explaining the CDF traces and the differences between the two tools that can make CDF traces (Citrix Scout and CDF Control). Bas discussed the requirements to use Scout, shows how to use Citrix Scout (check on the requirements) and how to handle the outcome of the trace (TMF files). Bas also touched the TaaS service. Last topic was about the clear text verbose logging.
After the lunch we continued with the first CYOS session provided by Robin Hobo with his presentation XenMobile 10: what’s new and how to set-up. He started with the new enhancements within XenMobile 10: one appliance, placement of the XenMobile appliance in the DMZ, followed by the requirements (firewall ports, hypervisor, SQL, Apple Push Notification Services Certificate for IOS devices, 4 IP addresses in DMZ, 2 public IP addresses, DNS Records, SSL certificates, NetScaler Gateway). Robin continued describing the set-up steps of XenMobile 10, including some tips and tricks. Robin also touched the infrastructure changes of the MAM communication with the NetScaler. Robin mentioned that the wizard can only be used if you have the enterprise license, otherwise you cannot do the full wizard. Robin mentions that the Android Scheduler is not created by default anymore, which is needed for automatic deployments. Robin also described the steps for creating a Device Policy. Nice addition is the possibility to add screenshots and FAQs per application. Robin also likes the support tools page. Also the consoles are touched as last topic.
Next we had our first foreign speaker of the day. Jim Moyle presented the session PVS Cache to RAM with overflow to disk analysis. Jim started explaining the different cache types on his own way and the characteristics of these types. Jim continued describing how the Cache to RAM with overflow to disk works (uses non paged memory pool, use .vhdx on the drivers), followed by the limitations (page file always on disk, risk of performance degradation when spilling over to disk, management and reporting limited to a tray icon). Next Jim described his test methodology (including a link to his article about IOmeter), followed discussing his results. The results are really interesting, overflowing to disk effects performance a lot. The performance when overflowing to disk is less good that the actual values of the storage. For persistent world you should not use it. Also the RAM was not giving fully back the memory when overflow to disk has taken place. Next topic Jim discussed how the data was divided between RAM and overflow to disk (First in First out in memory). The filter has some intelligent to bring back in memory when they are re-used. Jim recommends to use Device RAM with overflow to disk when you don’t have good storage, if storage is fast just use Client Hard drive as cache type. Jim summarized is that a good feature, but use the disk overflow only as a real back-up option.
After Jim the second CYOS was provided by Bader el Khalifi with his presentation PVS made simple. Bader started with the phrase of Albert Einstein Everything should be made as simple as possible, but not simpler. He started with an overview of which components he think are not needed. For that components he starts explaining why. He started with BDM (why?: auto created with PVS 7.x, Faster boot times, no TFTP required, no PXE required). Next Bader discussed local storage for PVS vDisks. To keep it simple Bader advices not to use vDisk versioning, the reactions in the audience were pretty different on this opinion. The underlying part that you should periodically build a new vDisk is agreed by everyone. The session ended with some discussions about MCS and PVS.
The last presentation was Helge Klein with the title How folder redirection impacts user experience and breaks applications. Helge started explaining how the API works for folder directions and how that impacts Folder Redirection by changing User Shell Folders in several ways. Helge continued with demo showing the impact of resource exhausting on a file server without and with Folder Redirection. After the demos the complete measurements were discussed by Helge of several scenarios. Next topic touched by Helge was SMB latency and launch redirection with different resource load on the file server and different scenarios (roaming profile, folder redirection and Citrix Profile Management). Helge also described the improvements in SMB 2.x in comparison with version 1.x. They measured application launch time with different SMB versions used on the file server. Helge ended the presentation with recommendations: apps ignoring the API breaks folder redirection, consistent file server performance throughout the day is essential, folder redirection is trading logon performance for user experience for the rest of the day, redirected folders are automatically made offline, monitor your file server performance, and think twice before redirecting AppData.
Last session of the day is traditionally also Geek Speak Live hosted by Barry Schiffer and Kees Baggerman. They built around it one theme: Does the Citrix product stack offers everything I need, where Barry and Kees picked several components Citrix have or have not a solution (profile, monitoring, hypervisor and storage) and discussed that with the audience.
We ended the day traditionally with some drinks and food, where the day was discussed by the attendees.