question

$$ANON_USER$$ avatar image
0 Votes"
$$ANON_USER$$ asked saldana-msft edited

mecm windows 10 upgrade provisioning mode

We have a large client base cir. 380k clients. MECM is 2010.

We are deploying a Windows 10 inplace upgrade from 1803/1809 to 1909 dec image and are starting to see a high volume of calls of devices that are no longer receiving policy therefore users are unable to see deployments.

After investigations it seems these devices have become stuck in provisioning mode, we are aware of the reg fix to remove them from provisioning mode or to just let the device un-provision itself after 48hrs but would like to know if anyone has developed a proactive way of identifying these devices that are stuck in provisioning mode opposed to being reactive to help reduce the impact on users and high call volumes?

Thanks

mem-cm-generalmem-cm-osd
5 |1600 characters needed characters left characters exceeded

Up to 10 attachments (including images) can be used with a maximum of 3.0 MiB each and 30.0 MiB total.

Jason-MSFT avatar image
0 Votes"
Jason-MSFT answered

It's not easy necessarily and will be time consuming, but I would start with root cause here of why they are left in this state as this is not normal, expected, or by design. Removing the agent from provisioning mode addresses a symptom and not the cause -- there may be other symptoms here as well that are not corrected by simply removing it from provisioning mode.

5 |1600 characters needed characters left characters exceeded

Up to 10 attachments (including images) can be used with a maximum of 3.0 MiB each and 30.0 MiB total.

$$ANON_USER$$ avatar image
0 Votes"
$$ANON_USER$$ answered

Hi Jason, yes agree.

We currently have MS assisting with the issues root cause and hopefully be able to provide us with possible fixes.

I am looking for a way of trying to be proactive rather than reactive so we can maybe identify some problematic devices i.e a good way to find devices that have been in provisioning mode for a long period of time or devices that seem to be stuck installing the TS etc.. in the hope we can attempt to use the possible fix to resolve the issue before it becomes an issue to the end user which results in calls being raised.

Just reaching out to see if anyone had found a way to do this, I know we wont be able to catch everything easily but if we can be proactive on some devices it is better then none.

5 |1600 characters needed characters left characters exceeded

Up to 10 attachments (including images) can be used with a maximum of 3.0 MiB each and 30.0 MiB total.

ColinFord-6663 avatar image
0 Votes"
ColinFord-6663 answered

Go take a look at Gary Blok's excellent WaaS task sequences. He has some fallback RunOnce keys in there that automatically remove the provisioning reg keys if the upgrade kills the TS before it can finish https://garytown.com/waas.

5 |1600 characters needed characters left characters exceeded

Up to 10 attachments (including images) can be used with a maximum of 3.0 MiB each and 30.0 MiB total.

GaryBlok avatar image
0 Votes"
GaryBlok answered

There is a script you can use on my blog post: https://garytown.com/configmgr-client-provisioning-mode

You should be able to use CMPivot or RunScripts to detect machines on ProvMode. I'm pretty sure you can leverage those even when the client is in provisioning mode, as it's more of a "push" than a "pull".

However, as Jason said, you need to find what is going on that is leaving the machines in provisioning mode. If the OS Upgrade step is completing successfully, it WILL automatically be pulled out of provisioning mode at that point.

Does this happen on all upgrades, on some? Specific models? What have the logs said on the machines that aren't coming out of provisioning mode, they must not be completing the TS successfully, are you getting Windows Setup Errors?

I'd suggest also servicing your media so it's completely updated, instead of using December media.

5 |1600 characters needed characters left characters exceeded

Up to 10 attachments (including images) can be used with a maximum of 3.0 MiB each and 30.0 MiB total.

$$ANON_USER$$ avatar image
0 Votes"
$$ANON_USER$$ answered GaryBlok commented

Hi Gary, yes we are looking at your scripts at the moment. :-)

It is only happening on a small percentage of devices but in a large environment which we have (380k clients) it turns into a large number of calls, the issue is local to the machine and as you know it could be due to different reasons, we have not been able to alleviate the issues before the problem with the TS occurs yet, that's tomorrows job.

Thank you everyone for your input so far.

· 1
5 |1600 characters needed characters left characters exceeded

Up to 10 attachments (including images) can be used with a maximum of 3.0 MiB each and 30.0 MiB total.

That's really interesting. We used to see this happen more in previous upgrades with older versions of CM, but didn't run into it at all this last upgrade (1809->1909 on CM 2006).

Typically the only time I see it happen is if the Setup Engine Fails in a way that the Task Sequence doesn't recover, leaving the machine in that state, and that's very rare now, the only way I can reproduce the issue is rebooting the machine during the task sequence while in the upgrade downlevel phase (once the Setup Engine has started and the machine is in Provisioning mode, but before the first automatic reboot).

That will cause the machine to be in a pretty bad state, but I've developed processes to resolve that situation leveraging scheduled tasks (all available in the WaaS Download on Garytown.com)

Good luck, I'll be interested to hear your continued progress on this.

0 Votes 0 ·