I'm not entirely sure if this is the best place to ask this, but I certainly figured that it wouldn't hurt (and hopefully perhaps someone has encountered something similar to it previously).

One of my colleagues recently encountered an issue when running Windows Updates on a 2008R2 machine that appeared to cause the CPU to throttle to 100%, which rendered the server nearly unusable at the time. He attempted the fix mentioned in this related discussion of disabling IPv6 on the server. Unfortunately, the issue returned two hours later (after a restart of the server) when updates were attempted once again.

Additional searches to try and find either the cause or a solution to the issue haven't been extremely fruitful and most of them are strictly Windows XP related (and generally older). I wasn't sure if this was an actual bug or some type of conflict, but any and all suggestions or recommendations would be appreciated as this has bugged him for quite some time now and he turned to me for assistance (as he doesn't generally reach out into the community much).

I would glad to provide any additional information or details as needed if it would help.

Thanks in advance,

Rion

Just as an update (after attempting various fixes, workarounds, etc.), it appears that this issue has been only reproducible on virtual machines.

Upon discovering this, we noticed that each of these virtual machines only had a single processor dedicated to them and during updates this processor would hit 100% CPU usage. However after applying two processors to one of the VMs, the throttling limited to 50% (i.e. just maxing out the single processor that the thread was running on), but the machine was at least responsive and usable during this time (unlike before).

Is Windows Update designed to run on machines using a single processor / core? As it certainly appears in this instance to be maxing out the one that it is running on. 

This may be completely unrelated to the initial issue at all, but it was just an idea that I had looked into as the other fixes had not been successful.

Another follow up as things are progressing.

It appears this issue only occurs on machines with some flavor of SQL Server installed on them. When Windows Update is run, the CPU usage will skyrocket into the upper 90% range and performance will obviously become sluggish during the course of the updates. This issue doesn't appear to occur at all on machines without SQL and we are currently just throwing two processors per via to mitigate the issue for the time being (thus CPU usage will peak at 50%, but the system will still be responsive).

My question is why would this behavior be occurring? Does Windows Update perform any differently when SQL Server is installed or is there anything going on behind the scenes that might explain this behavior?

Hi,

We are looking into this issue and will reply when we get progress. Thank you for your

Thanks Shaon, I appreciate it.

Hi Rion,

Sorry for the delay in reply. It takes us some time on discussion.

Above all, there is no limitation about Windows Update working on a single processor system. However Windows Update could take high system resource when it is downloading and installing updates. If 2 processors could reduce the CPU usage to 50%, it should not be an issue but just a "no enough system resource" situation. 

My teammate suggest me to collect these information from you:

When did this issue first happen? Have you done any configuration before that?
Where did you detect the update from? Windows update or WSUS?