We installed a Proxmox server for a client, and recently updated to it to ProxMox V3.4
The specification of the bare metal computer is -
- 2 x Intel Xeon E5-2650 Processor 2.0GHz 20MB Cache 8 Cores
- 32GB DDR3 Registered ECC RAM
- Intel Hardware SATA RAID Controller with battery backup
- 2 x Seagate 2TB 3.5” Enterprise Mission Critical SATA Drives
- 2 x Seagate 5TB 3.5” Enterprise Mission Critical SATA Drives
Proxmox and the VM installed on 1 of the 2TB HDD.
This is the only virtual on the Machine.
The VM is 800GB Running Win Server 2008R2, all updates loaded.
Partitioned Op-Sys and programs 55GB, 20GB free. Data 745GB, 155 GB free.
We used Virtio drivers for Network, Drives and Memory (1GB min 32GB max)
We do backup's to the other 2TB and the 2 x 5TB drives.
The VM is currently configured to act both as a production server, and after a VM copy, a test/development server with the ability to restore the earlier working production environment.
The reason we chose this scenario was the client was unwilling to outlay more cash for the licensing of a complete set of applications for the test/development server, and the applications cannot be copied to another server as they are hardware locked to a single server, and the application provider is unwilling or unable to provide a application set suitable for a test/development server.
The install proceeded well, all was going OK, installed Windows Server 2008R2, did updates and patches, added to the existing domain.
Applications and programs installed and initial test runs completed.
The server was moved into full production, and they use multiple RDP logins (between 2 to 10)
The applications run like a dog, very slowly, if at all. We are talking get a coffee, from the cafe across the street.
.
The Processor, in the Proxmox interface the CPU usage jumps to 85% to 100%.
Similar when looking at the VM CPU usage in the Proxmox interface.
Looking at the CPU usage from within the VM, under Windows it shows average of around 60%.
There is a variation in individual core usage.
During a production run, when the server load is running as described above, then we get the following ping times.
When we ping from a local pc to the Proxmox box bare metal network IP address - average ping times of <1ms
When we ping from a local pc to the Windows VM virtio network IP address - the ping times vary from <1ms to 60ms sometimes higher.
When the server cpu load drops - after the production runs have either finished or been terminated for the day, the ping times to the Windows Virtio network IP has ping times that are <1ms
Both local and remote users are unable to use the system as it now stands.
As you can imagine the remote users find it very, very difficult to use the system.
Not sure where to turn next, your ideas / suggestions / help would be appreciated.
The specification of the bare metal computer is -
- 2 x Intel Xeon E5-2650 Processor 2.0GHz 20MB Cache 8 Cores
- 32GB DDR3 Registered ECC RAM
- Intel Hardware SATA RAID Controller with battery backup
- 2 x Seagate 2TB 3.5” Enterprise Mission Critical SATA Drives
- 2 x Seagate 5TB 3.5” Enterprise Mission Critical SATA Drives
Proxmox and the VM installed on 1 of the 2TB HDD.
This is the only virtual on the Machine.
The VM is 800GB Running Win Server 2008R2, all updates loaded.
Partitioned Op-Sys and programs 55GB, 20GB free. Data 745GB, 155 GB free.
We used Virtio drivers for Network, Drives and Memory (1GB min 32GB max)
We do backup's to the other 2TB and the 2 x 5TB drives.
The VM is currently configured to act both as a production server, and after a VM copy, a test/development server with the ability to restore the earlier working production environment.
The reason we chose this scenario was the client was unwilling to outlay more cash for the licensing of a complete set of applications for the test/development server, and the applications cannot be copied to another server as they are hardware locked to a single server, and the application provider is unwilling or unable to provide a application set suitable for a test/development server.
The install proceeded well, all was going OK, installed Windows Server 2008R2, did updates and patches, added to the existing domain.
Applications and programs installed and initial test runs completed.
The server was moved into full production, and they use multiple RDP logins (between 2 to 10)
The applications run like a dog, very slowly, if at all. We are talking get a coffee, from the cafe across the street.
.
The Processor, in the Proxmox interface the CPU usage jumps to 85% to 100%.
Similar when looking at the VM CPU usage in the Proxmox interface.
Looking at the CPU usage from within the VM, under Windows it shows average of around 60%.
There is a variation in individual core usage.
During a production run, when the server load is running as described above, then we get the following ping times.
When we ping from a local pc to the Proxmox box bare metal network IP address - average ping times of <1ms
When we ping from a local pc to the Windows VM virtio network IP address - the ping times vary from <1ms to 60ms sometimes higher.
When the server cpu load drops - after the production runs have either finished or been terminated for the day, the ping times to the Windows Virtio network IP has ping times that are <1ms
Both local and remote users are unable to use the system as it now stands.
As you can imagine the remote users find it very, very difficult to use the system.
Not sure where to turn next, your ideas / suggestions / help would be appreciated.