Windows Server 2008 R2 consuming all available RAM for CACHE
I am running a windows server 2008 R2 standard as a host for 6 virtual machines. The server has 2 x quad core XEONS and 32GB of Ram.
After I restart the host, I can boot all the virtual machines boot and consume the expected amount of RAM and for a few hours cache RAM remains low when viewed from TASK Manager. If I log in 6 to 10 hours after the host is restarted and all VM's are running, the HOST has consumed all remaining RAM (as much as 12GB) into CACHE and leaving ZERO FREE memory.
At first I thought this might be a Backup Program I was running, so I uninstalled it, and other than the base services for the HOST, I have no other applications running on the HOST. I have no idea why the server is consuming ram and I'm not sure if this is or can be affecting the performance of the VM's. Any thoughts on how to manage the CACHE memory or even how to clear it would be appreciated.
After I restart the host, I can boot all the virtual machines boot and consume the expected amount of RAM and for a few hours cache RAM remains low when viewed from TASK Manager. If I log in 6 to 10 hours after the host is restarted and all VM's are running, the HOST has consumed all remaining RAM (as much as 12GB) into CACHE and leaving ZERO FREE memory.
At first I thought this might be a Backup Program I was running, so I uninstalled it, and other than the base services for the HOST, I have no other applications running on the HOST. I have no idea why the server is consuming ram and I'm not sure if this is or can be affecting the performance of the VM's. Any thoughts on how to manage the CACHE memory or even how to clear it would be appreciated.
Who is Participating?
Update, I installed Microsoft Security Essentials a ran a virus scan to see if it was a bug, the scan came back clean. Attached is an image of the results of RAMMAP: the memory being consumed is in MAPPED FILE
Cache memory is not necessarily/technically consumed memory. It can be allocated to processes requesting memory. If you want to see how much memory is actually available to processes, start the resource monitor (Start->Programs->Accessor
Hardware Reserved - video memory, BIOS, HW resources, etc
In Use - claimed by OS or processes
Modified - memory that has data that would have to be flushed to disk before released for swapping
Standby - memory released but still linked to processes, can be reallocated to processes
Free - not yet allocated memory
The "Available" entry at the bottom of the graph is essentially what the OS can provide to requesting processes.
Here's an article that provides a lot more details
Hardware Reserved - video memory, BIOS, HW resources, etc
In Use - claimed by OS or processes
Modified - memory that has data that would have to be flushed to disk before released for swapping
Standby - memory released but still linked to processes, can be reallocated to processes
Free - not yet allocated memory
The "Available" entry at the bottom of the graph is essentially what the OS can provide to requesting processes.
Here's an article that provides a lot more details
Update: I have removed 12 GB of RAM since that was the lat thing I did before I noticed this problem. I am now running 24GB ram
Thanks gt2847c:
Attached is a copy of the resource monitor stats: Based on what you are saying, the "10891 MB Available" which includes 777MB of FREE ram is actually available for other processes and services, so then I should not be concerned about performance issues like Network connections or disk writes. I will read the article.
Thank you.
ResourceMonitor.jpg
Attached is a copy of the resource monitor stats: Based on what you are saying, the "10891 MB Available" which includes 777MB of FREE ram is actually available for other processes and services, so then I should not be concerned about performance issues like Network connections or disk writes. I will read the article.
Thank you.
ResourceMonitor.jpg
Question has a verified solution.
Are you are experiencing a similar issue? Get a personalized answer when you ask a related question.
Have a better answer? Share it in a comment.
All Courses
From novice to tech pro — start learning today.
Memory tab - Hard Faults/sec and Commit Charge. These will go high when you run into memory contention (need more RAM on the system)
Disk tab - Queue Length. This is how many things are waiting disk IO. If this number gets high, your system is IO constrained on storage. If this is a production system with performance requirements, you may have to rearrange VMs, storage setup and such to lower the queue length. Adding another storage array, moving VM images to other disks, etc. can help lower IO contention and improve performance.
Network - Interface utilization. You need to make sure you're not over-committing your network interface for your VMs. Too much contention here will cause performance issues. If you have high utilization on a given NIC, consider adding another NIC and rearranging your VMs to spread the load around.
VM systems generally require tuning to make sure you're providing good performance. Keeping an eye on the trending of the various subsystems on your host machine can help you prevent getting surprised by performance issues...