Jump to content


anyweb

Root Admin
  • Content Count

    8,456
  • Joined

  • Last visited

  • Days Won

    332

anyweb last won the day on October 14

anyweb had the most liked content!

Community Reputation

475 Excellent

About anyweb

  • Rank
    Administrator
  • Birthday 11/24/1966

Contact Methods

  • Website URL
    http://www.niallbrady.com

Profile Information

  • Gender
    Male
  • Location
    Sweden
  • Interests
    Modern management of devices with Microsoft Intune and System Center Configuration Manager

Recent Profile Visitors

4,147,241 profile views
  1. ok first things first, the SMSTSPostAction variable is for use in operating system deployment task sequences to do an action after the task sequence has completed, are you planning on installing SEP as part of a task sequence ? if so use an Install Application step instead of Install Package if that's easier, or even run PowerShell script... but first, if you really want to install the app just using a powershell script then test the script on a virtual machine standalone, outside of a task sequence
  2. I'd suggest you test Powershell Application Deployment Toolkit, it's what most Enterprises use to install apps in ConfigMgr today
  3. Yup, i've just tested the update function within Smoothwall, it works and the updates are right up to this year, if you get any complaint from the Smoothwall about not being able to update then reboot the Smoothwall and try again
  4. interesting, it also seems that Microsoft is recommending the Monthly Enterprise Channel for office (while at the same time recommending Current Channel, see above...), which of course solves the problem but lacks features, another approach is to pay HP more money to ship devices without Office in the Corporate Image, that works too
  5. Introduction This blog post is long over due and I've been asked to do it many times, today, I finally got around to writing it. Sorry for the delay. Using Hyper-v labs to test all the new Endpoint Manager releases that Microsoft produces means you probably want to have multiple labs on the go at any time, one for Current Branch, one for Technical Preview, and another for testing out PKI or some other functionality. Having internet access in those labs is also usually a requirement, and keeping them separate from each other is also important. Using a Smoothwall to control internet into (and out of) each hyper v lab gives you control over when that lab gets access to the internet or not. I've used a Linux based firewall solution called Smoothwall, and it works great, so much so that I have one Smoothwall virtual machine for each lab. The Smoothwall allows me to configure individual ports to virtual machines within each lab and lots more functionality. Some Linux experience helps, but it's really easy to use. I've been asked several times over the years how I setup my Smoothwall and never got around to creating a guide for it, so here goes. Are there other ways of doing this, of course, but this is the way I do it. Step 1. Get the ISO Download the Smoothwall express i586 ISO from here. It's a 196MB download. http://sourceforge.net/projects/smoothwall/files/SmoothWall/3.1/Express-3.1-i586.iso/download Step 2. Create network switches In hyper-v manager, create 2 network switches, one Private Network Switch for your lab (we'll name it #11 in this example) and one switch connected to a physical network card (either WI-Fi based or Ethernet) to share the internet into the lab, we'll call that ICS for Internet Connection Sharing. Below is the LAB network switch, it's private so computers within this individual lab can talk to each other but cannot talk to other labs on my host. Below is the ICS switch, note that I do not allow the management operating system to share this network adapter. That can cause all sorts of problems in the host operating system, so don't select it. Also, this is a WI-Fi nic but it could be an Ethernet adapter, as long as it has internet in, it's good to go. Step 3. Create a new virtual machine In this step I'll use a PowerShell script to create a new virtual machine to host Smoothwall, the important thing to remember is that the virtual machine must be Legacy (type 1) and the network card must also be a Legacy Network Adapter. Here's an example for my lab #11. Note: You only need 256 MB of ram for this virtual machine so either change the script or modify the virtual machine settings later. Step 4. Add additional Legacy Network Adapter In the virtual machine just created, add a new Legacy network adapter and point it to the ICS virtual switch. Step 5. Install Smoothwall Attach the ISO to the CD rom drive in the Smoothwall virtual machine, change the boot order to CD and boot. Choose to install Smoothwall Express. click Ok to the welcome and click OK to the file system preparation. Click OK to erase. and it's done. When prompted to restore the configuration choose No. Select your keyboard layout, Select timezone Give it a hostname half open is fine... for Network Configuration Type choose GREEN + RED as shown below Use TAB to move to Done, think of the two colours as follows: GREEN: Lan RED: Internet then click OK to no green interface assigned, and next, select Assign network cards by clicking on Card Assignments Click ok when prompted to set them you'll be shown the 2 nics identified by MAC address in the following screens, so you can see which nic is assigned to which colour. after that you'll be informed that all cards are successfully allocated, move to Address settings... As Green is the lab IP of our smoothwall, I normally follow the ip address settings within each lab as follows 192.168.X.1 where X is the number of my lab, so this is lab #11 therefore I always use the following for each lab, the only thing that changes is the number replacing X DC01=DNS/DHCP/Domain Controller = 192.168.11.1 CM01 = Endpoint Manager = 192.168.11.2 Smoothwall = Linux firewall = 192.168.11.199 click ok and here I configure the IP address to match the ip address range shown above for my #11 lab. click ok, select the RED interface I normally use DHCP on RED and get a separate IP from my home network click ok, then Done when complete. and tab again to Done (or configure DNS if needed) Click Finished for Web Proxy Next set the Admin password set the root password click ok when done ! Step 6. verify after the reboot, login as root, it is case sensitive so at the prompt below type root and enter the password when prompted. after logging in, type the following to list the ip addresses assigned to your smoothwall. ifconfig you will see output something like so...it probably scrolled off screen, that's ok. In the example above: eth1 is the smoothwall RED interface and has my local network ip (from my Wi-Fi router), eth0 is the smoothwall GREEN interface and has the ip address I manually assigned it namely 192.168.11.199 You can issue the following commands: ifconfig eth0 ifconfig eth1 to list each nic individually. Next, try pinging some address: control +c to cancel, it works ! On a Windows virtual machine in the lab you are providing internet access (and routing) to, your network settings should be configured like so...pointing your Gateway to the smoothwall and DNS to the Domain Controller, this can be configured via DHCP server settings on the DC. a quick ping to verify internet works Step 7. Optionally configure port forwarding On a Windows machine, open a web browser and browse to the ip address of your Smoothwall and include port 441 like so https://192.168.11.199:441 you'll probably get a warning, it's safe to ignore Click Advanced to continue...for username and password use the Admin user you created for web configuration In this example I'm forwarding port 80 from my external internet connection to my internal lab, specifically the web server in lab #11. Job done, I hope you found this useful ! If you'd prefer to watch a video of this then see here cheers niall
  6. we've seen the issue is back today, so we've switched to the following version, it works.
  7. Introduction I'm writing this post to catalog problems I recently faced while doing Windows Autopilot installations, we use a very slim enrollment Status Page (ESP) configuration with only one app marked as required (Microsoft Edge Chromium). The remaining applications are a mix of Win32 apps and the all important Microsoft Office 365 suite. This suite was configured with the following apps. Excel, OneDrive Desktop, OneNote, Outlook, PowerPoint, Teams, Word This worked well since about mid-February 2020. However, starting week 46, I started noticing the following error on newly delivered HP computers with Windows 10 version 1909 (and some office components preinstalled) after clicking the Microsoft Outlook icon soon after Windows Autopilot had completed. It looks ominous. Clicking OK and trying again, the problem looked even worse, you'd see something like this, outlook prompting you to choose a Profile. followed by a Script error, stating Class not registered on Line 278. Choose Yes or No had pretty much the same effect, Outlook was well and truly broken. If you clicked on the Account Information in Outlook you'd see something like this. A Metered connection warning followed by a Upgrade in Progress warning. The first is definitely a red-herring as the Network card was not in a state that was metered. When checking the version of office installed on affected machines I could see it was as shown below, Office version 2011. The interesting thing to note is that Computers that had a clean image of Windows 10 version 1909 with no Office installed previously did not exhibit this problem, it only affected factory image(s) of HP computers with Windows 10 version 1909 factory image and those images come with a version of Microsoft Office installed (in the Nordics), namely Microsoft Office 365 ProPlus version 1908 (Build 11929.20394). Troubleshooting Based on the above I knew that clean installs of Windows 10 1909 did not have the issue (even though they subsequently got the Office 2011 version installed before the user logged on). I initially suspected that security software or a device configuration profile were to blame, and went through the time consuming task of excluding a computer from each profile, and then resetting it to verify the behavior. Excluding a device from an assignment takes precedence over including a device so it was a good way of testing Windows Autopilot without certain settings or configurations, to rule them out. Below you can see I've excluded a group (containing my test device) from a Device Configuration profile, to verify if that was the issue. trying this didn't help, but it at least ruled out the following from being part of the problem. Device Configuration Profiles Win32 based Security based apps (such as Azure Information protection, Crowdstrike, Symantec DLP) Armed with that knowledge I recreated the Office Suite settings in my own test tenant, and ran a Windows Autopilot build, to my surprise the HP failed starting outlook the exact same way as in Production, so that completely ruled out everything other than the version of Office installed on the HP. Next I turned to logging options within Office/Outlook to see if that would help, but in reality it just generated .ETL files that I'm still analyzing in order to root-cause this issue. The breakthrough came when looking at the settings of the Office suite in Endpoint Manager. The version of Office that gets installed is based on your settings in the Office Suite, and we had been using these settings without problem since February 2020. The really important bit was the update channel, shown below. The update channel we were using was Current Channel (Preview). You can get details of the update channels here. According to Microsoft: ... three primary update channels: Current Channel Monthly Enterprise Channel Semi-Annual Enterprise Channel We recommend Current Channel, because it provides your users with the newest Office features as soon as they are ready. But what is the difference between Current Channel and Current Channel (Preview). According to Microsoft: To become familiar with the new features coming in the next feature release of Current Channel, we recommend that you use Current Channel (Preview). There isn’t a set release schedule for Current Channel (Preview). In general, a new version of Current Channel (Preview) with new features is released at least a week or more before that new version is released to Current Channel. There might be several releases of Current Channel (Preview), with non-security updates, before that version is released to Current Channel. You should deploy Current Channel (Preview) to a small, representative sample of users in your organization. This can help you identify any possible issues for your organization before those new features are released more broadly to your users that have Current Channel. We also encourage you to use Current Channel (Preview) so that you can identify any possible issues that you want us to fix before that version is released to Current Channel. This can help reduce the number of non-security updates that are needed for Current Channel. And this pretty much matched what we were doing, so now that we had this knowledge, but still had no root-cause for the Outlook (and Word/Excel issues). The Resolution I decided to change the Update Channel from Current Channel (Preview) to Current Channel. This decision was based on the fact that the Preview channel may contain changes that are incompatible with our image in some way, which is odd because we are using the factory installed HP image. Once I made the change, and re-tested Windows Autopilot the difference was clear. Now Outlook worked as expected without issue (and Word/Excel issues disappeared also), however the version of Office installed was Version 2010 instead of Version 2011 that we got in the Current Channel (Preview). This didn't matter too much but of course it meant that some cool new cloud friendly features in Version 2011 were now no longer available on newly installed Windows Autopilot machines. The versioning used by Microsoft for Office is somewhat confusing, in the Office Account screen you'll see the version info, here you can see it's listed as Current Channel, Version 2010 (the version without the problem). So version 2010 relates to year 20, month 10, or the October release of Office 365. That would of course mean that version 2011 is the November release. Note: we've noticed that HP's corporate ready image includes an office version that is released before the OS version was released, so for example if you get Windows 10 version 1909, then you'll get the Office version released approximately one month before that (Office version 1908). Likewise if you got Windows 10 version 2004, you should get Office version 2003. Summary Sometimes living on the edge means you will fall over. I know that changing from Current Channel (Preview) to Current Channel might only delay the problem until the Current Channel update channel migrates to the new version of office next month, so we may actually encounter this problem again, and soon. So to conclude, if any of you have come across this exact issue (I have searched and found similar problems with "Library not registered", but the advice within them didn't apply here), then please get in touch with me. In the meantime I will look through the gathered ETL traces to see if they provide any clue as to why Office was so broken on these new devices in order to root-cause the problem. Links used in this blog post Github script, Metered - https://gist.github.com/nijave/d657fb4cdb518286942f6c2dd933b472 Update Channels - https://docs.microsoft.com/en-us/deployoffice/overview-update-channels Office Versions - https://docs.microsoft.com/en-us/officeupdates/current-channel
  8. get-windowsupdatelog on the machine in question will allow you to see where the updates is being downloaded from
  9. hi Scott and welcome the first thing i'd suggest you do is grab the smsts*.log files on one of the affected computers and zip them up and upload here so we can take a look, that will hopefully reveal what went wrong any additional info you can add, such as hardware that was affected and whether or not there's any changes to BIOS/UEFI setup in the task sequence ?
  10. yup that should improve things try it however are you sure your computers are only getting updates from ConfigMgr and not somewhere else ?
  11. are you referring to the TPM hash ? if so read here TPM password hash Previous MBAM clients don't upload the TPM password hash to Configuration Manager. The client only uploads the TPM password hash once. If you need to migrate this information to the Configuration Manager recovery service, clear the TPM on the device. After it restarts, it will upload the new TPM password hash to the recovery service. Uploading of the TPM password hash mainly pertains to versions of Windows prior to Windows 10. Windows 10 by default does not save the TPM password hash so therefore does not normally upload the TPM password hash. For more information, see About the TPM owner password.
  12. while it's not a report, have you looked at the Cloud Management overview in the ConfigMgr console ? for more details about monitoring the clients to your CMG and and traffic involved see https://docs.microsoft.com/en-us/mem/configmgr/core/clients/manage/cmg/monitor-clients-cloud-management-gateway to figure out how much it costs see
  13. anyweb

    Hello from Indiana!

    ah ok, well it's out of support and you'll need to ugprade it to a supported build before you decide to do anything with it, that would be your first task, to setup a plan for the upgrade, it's easy, and you'll gain access to lots of new features by doing so (as well as remain supported) cheers niall
×
×
  • Create New...