VMware Cloud Community
bwalsh1
Contributor
Contributor
Jump to solution

Added RAM To Host, Guest CPU Utilization Goes Up.

I upgraded the RAM on my ESXi host from 32GB to 128GB.  After power back on and making no changes to any guests the CPU of my guest workstations is extremely high.  I have 3 Windows servers that hover around 70 percent CPU utilization.  They normally hover around 5 percent.  Any ideas?  Thanks in advance.

0 Kudos
1 Solution

Accepted Solutions
bwalsh1
Contributor
Contributor
Jump to solution

After 3 days I solved this issue.  Turns out it was never a VMWare issue.  The server (Lenovo ThinkServer RD650) had 32GB of memory and never had an issue.  I removed the 32GB of RAM and inserted 128GB of RAM.  My 4 guest servers, all Windows servers, ran very slow.  Normal CPU utilization of them was an average of 10 percent.  After the RAM upgrade they idled at 60 CPU percent utilization.  Any work they did would max out the processors.  All the while the host processor was around 5 to 10 percent utilization.  I immediately shut down the host and put the 32GB of RAM back in thinking it was bad or improper RAM.  When I got all the guests running on the original 32GB they were still acting up just as bad as with the 128GB.  For 2 days I checked and double checked settings and could not find an issue.  Then I noticed an alert for 4 of the fans not working.  Previous to the RAM upgrade I knew there was no alarm because I checked everything.  I scheduled downtime (Windows Update destroyed my time allotted) and confirmed that every fan was good.  I re seated the fans back in their bays and powered everything up.  VMware showed the fans as running and the problem was solved.  Everything was fast like it used to be.  Even though the processor temp was in range the processor was downgraded because of the lack of fans.  Hopefully this post helps somebody else out.

Thank you for all the replies.  I read up on all of your suggestions and learned a ton about how to monitor and configure VMware ESXi.

View solution in original post

5 Replies
maksym007
Expert
Expert
Jump to solution

Are you sure that you have appropriately mounted RAM Modules? 

 

How many CPUs are physically installed on the Server? For each CPU - you have dedicated RAM Modules slots. On the Server cover - you have a Mainboard plan for which RAM to which CPU is assigned. Check this first. 

0 Kudos
bwalsh1
Contributor
Contributor
Jump to solution

Thank you for the reply.  The server is a Lenovo RD650 with one processor installed.  I did follow the instructions for installing the 4 DIMMS in the appropriate slots.  After further investigating this seems to be a memory speed issue.  The original DIMMS are 2133MHZ, the ones recommended on the website where I purchased them are  3200MHZ.  The server still booted but ran excessively slow.  I re installed the original RAM and it does run better but, not like it used to.  Once again, thank you for the reply.

0 Kudos
maksym007
Expert
Expert
Jump to solution

Thx for your feedback. Good to know that you have identified the source of the problem.

0 Kudos
Tibmeister
Expert
Expert
Jump to solution

Speed mismatch can cause a whole host of issues.  Check some of the standard stuff like Co-Stop, CPU Ready, etc.  Also, make sure the CPU can support that much RAM, and the speeds that it supports, along with the technology.  Lastly, make sure you aren't messing with vNUMA, stick with the default 1 CPU per Core setting for the VM CPU layout.

0 Kudos
bwalsh1
Contributor
Contributor
Jump to solution

After 3 days I solved this issue.  Turns out it was never a VMWare issue.  The server (Lenovo ThinkServer RD650) had 32GB of memory and never had an issue.  I removed the 32GB of RAM and inserted 128GB of RAM.  My 4 guest servers, all Windows servers, ran very slow.  Normal CPU utilization of them was an average of 10 percent.  After the RAM upgrade they idled at 60 CPU percent utilization.  Any work they did would max out the processors.  All the while the host processor was around 5 to 10 percent utilization.  I immediately shut down the host and put the 32GB of RAM back in thinking it was bad or improper RAM.  When I got all the guests running on the original 32GB they were still acting up just as bad as with the 128GB.  For 2 days I checked and double checked settings and could not find an issue.  Then I noticed an alert for 4 of the fans not working.  Previous to the RAM upgrade I knew there was no alarm because I checked everything.  I scheduled downtime (Windows Update destroyed my time allotted) and confirmed that every fan was good.  I re seated the fans back in their bays and powered everything up.  VMware showed the fans as running and the problem was solved.  Everything was fast like it used to be.  Even though the processor temp was in range the processor was downgraded because of the lack of fans.  Hopefully this post helps somebody else out.

Thank you for all the replies.  I read up on all of your suggestions and learned a ton about how to monitor and configure VMware ESXi.