Hey everyone, I'm facing some serious performance issues with our terminal server that's impacting our users. Here's a bit about the setup: we have one physical server running Windows Server 2022 Standard, hosting a couple of VMs: one for Active Directory and DNS, and another for terminal services along with some line-of-business (LOB) applications. The physical server has a Xeon Silver 4316 processor, 128GB RAM, and a 40TB RAID10 HDD setup (about 20TB usable). The terminal server VM is configured with 96GB of RAM and 12 vCPUs with around 14TB of storage, while the DC VM has 4GB of RAM and 4 vCPUs. At any time, we have about 5-10 users logged in, but their experience remains sluggish, with noticeable long load times and frequent application crashes across multiple software, including Outlook and Acrobat. I've already tried several things like reinstalling Office and Acrobat, adding paths to antivirus exceptions, scheduling nightly server reboots, and tweaking backup policies, but the issues persist. I monitor CPU and RAM usage closely and they seem fine, but the disk activity frequently spikes. My next plan is to contact our LOB vendor for further help, but I'm hoping someone here might have additional tips or strategies to improve performance. Any insights would really help!
4 Answers
Definitely check those event logs and make sure all drivers and firmware are updated. You mentioned that your server is already up-to-date, which is great, but it’s still worth a double-check to rule out anything there.
Also, keep an eye on CPU ready time and disk latency if you can. Your figures seem decent, but any spikes might be contributing to the performance issues. You definitely want to minimize latency at the hypervisor level.
I appreciate that! I'll monitor those metrics more closely. It's a good point that a sudden spike could be the culprit.
It sounds like your biggest bottleneck might be the HDD speed. Those 7.2k RPM drives in RAID10 can struggle with the random I/O loads from Office, Acrobat, and your SQL-based apps. You might want to consider upgrading to SSDs, especially for the OS drives on your VMs. It'll greatly improve access times.
Yeah, SSDs can make a huge difference! We experienced significant performance improvements once we moved to SSDs for our remote desktop servers.
Just a suggestion, but is there a chance you could implement user profile redirection? Storing profiles locally on the terminal server might be contributing to the slowdown, especially with multiple users accessing it.
That's something I haven't considered! I'll look into profile redirection to see if it would help optimize performance.
Thanks for the reminder! I did update everything via Lenovo's XClarity, but I'll take another look at the logs for any potential issues.