Hello, I use windows 2025 on RS 2000, at first it works just fine, but when deleting some files, the usage is stuck.
after some time the disk usage goes above my plan and leads to freezing of my server.
Hello, I use windows 2025 on RS 2000, at first it works just fine, but when deleting some files, the usage is stuck.
after some time the disk usage goes above my plan and leads to freezing of my server.
You might be suffering from the same issue as discussed here (in German) by me and another customer.
The problem is, when disk is full or near-full the I/O performance drops to nearly zero and the server hangs completely.
It seems like customers with disk encryption are suffering more often (because disk is always full from the view of the host).
If you state that usage goes above your plan (your 2nd screenshot is not showing that) it could be the case that you are running into this issue as well.
You can temporarily fix this problem by shutting your server down and start it via SCP (a "normal" reboot does not help). Unfortunately, the problem will reoccur about 24 to 48 hours later.
but when deleting some files, the usage is stuck.
That is because the blocks are still marked as in-use by the filesystem (NTFS). This can be solved by defragmenting the drive every once in a while, or activating TRIM.
after some time the disk usage goes above my plan and leads to freezing of my server.
Do you have snapshots?
You might be suffering from the same issue as discussed here (in German) by me and another customer.
The problem is, when disk is full or near-full the I/O performance drops to nearly zero and the server hangs completely.
It seems like customers with disk encryption are suffering more often (because disk is always full from the view of the host).
If you state that usage goes above your plan (your 2nd screenshot is not showing that) it could be the case that you are running into this issue as well.
You can temporarily fix this problem by shutting your server down and start it via SCP (a "normal" reboot does not help). Unfortunately, the problem will reoccur about 24 to 48 hours later.
Yes, exactly, but I had to use optimize and that takes to much time.
I am almost at my plan limit, I will try the SCP reboot and see what happens.
That is because the blocks are still marked as in-use by the filesystem (NTFS). This can be solved by defragmenting the drive every once in a while, or activating TRIM.
Do you have snapshots?
What is TRIM?
Also I had one snapshots, but I deleted it when tried to optimize my drive.
I thought it was something in netcup control panel.
Looks like it's enabled by default, then trim won't fix the issue.
Well, if you are very close to your plan limit you may run into the issue I talked about earlier.
If you are using Bitlocker (encryption) it is even more likely to occur.
However, we hope the netcup team resolves that issue soon.
Anyhow, what could help in the meantime is:
- do not fill up your SSD space to the last byte
- shutdown your host and start using SCP if the server runs into I/O problems
- Did you install qemu-guest-agent? If your drive is not encrypted I think it tells the host about freed blocks if you delete files...
I just received a reply to my support ticket that the problem should be solved now and should not occur in the future anymore.
However, they say you should run a storage optimization once now such that it is resolved.
Alles anzeigenWell, if you are very close to your plan limit you may run into the issue I talked about earlier.
If you are using Bitlocker (encryption) it is even more likely to occur.
However, we hope the netcup team resolves that issue soon.
Anyhow, what could help in the meantime is:
- do not fill up your SSD space to the last byte
- shutdown your host and start using SCP if the server runs into I/O problems
- Did you install qemu-guest-agent? If your drive is not encrypted I think it tells the host about freed blocks if you delete files...
- I don't fill it 100%, I always keep 100GB free space.
- I don't use Bitlocker
- I did install qemu-guest-agent.
I just received a reply to my support ticket that the problem should be solved now and should not occur in the future anymore.
However, they say you should run a storage optimization once now such that it is resolved.
Thanks, I will shutdown the server once and optimize my storage once, then I will see if it works and report back here in the thread.
When the server os working I don't see the message
But when I shut it down, I see the message
I have installed qemu-guest-agent from the DVD virtio-win-0.1.271_20250711.iso
I wonder if there is a bug in netcup detection of qemu-guest-agent
Also the reported disk usage is wrong even after installing qemu-guest-agent and using "Storage optimization" feature
I wonder if there is a bug in netcup detection of qemu-guest-agent
The message is simpy misleading; it looks like it is simply the result of a failed test executed at the wrong time – obviously, once you turn off your KVM instance, qemu-guest-agent is unable to communicate with the host. I guess some kind of log analysis that would tell you something like "qemu-guest-agent was detected in the past, but the instance is currently turned off so it's impossible to run a test right now" (and, of course, an explicit result of said test when the instance is up and running) would be preferred by many users …
The issue you are describing with qemu-guest-agent is normal. In case the server is running and the message disappears, everything is fine.
The difference of storage being reported by windows and what is reported by SCP is normal. There is no problem if the storage of SCP is higher. This is simply due to the fact that if you delete a file on windows, the blocks are not freed by NTFS, just marked as usable.You are still able to use the full 1TiB.
If you don't trust it, you could possibly do the following experiment:
* Download the tool sdelete.exe from Microsoft. (https://learn.microsoft.com/de-de/sysinter…wnloads/sdelete). Run it on your drive C. Use the parameter "-z". This really deletes data from empty blocks
* After using the tool, SCP should show you a usage of 100% or another very high usage
* Shut your server down and perform a storage optimization
* Afterwards, SCP should show you the same percentage used as Windows reports.
Please note that this experiment doesn't help with performace. Over the time, if you create and delete files in Windows again, SCP will again show larger numbers. Also, this experiment would not work if encryption/Bitlocker is activated.
I suspect that the storage problems you experienced earlier had to do with the bug that your storage grows over your plan limit. This should be fixed, at least on my side it did not occur anymore within the last 24 hours.
The message is simpy misleading; it looks like it is simply the result of a failed test executed at the wrong time – obviously, once you turn off your KVM instance, qemu-guest-agent is unable to communicate with the host. I guess some kind of log analysis that would tell you something like "qemu-guest-agent was detected in the past, but the instance is currently turned off so it's impossible to run a test right now" (and, of course, an explicit result of said test when the instance is up and running) would be preferred by many users …
That makes sense, thank you, I will consider qemu-guest-agent installed and will never worry about it.
Alles anzeigenThe issue you are describing with qemu-guest-agent is normal. In case the server is running and the message disappears, everything is fine.
The difference of storage being reported by windows and what is reported by SCP is normal. There is no problem if the storage of SCP is higher. This is simply due to the fact that if you delete a file on windows, the blocks are not freed by NTFS, just marked as usable.You are still able to use the full 1TiB.
If you don't trust it, you could possibly do the following experiment:
* Download the tool sdelete.exe from Microsoft. (https://learn.microsoft.com/de-de/sysinter…wnloads/sdelete). Run it on your drive C. Use the parameter "-z". This really deletes data from empty blocks
* After using the tool, SCP should show you a usage of 100% or another very high usage
* Shut your server down and perform a storage optimization
* Afterwards, SCP should show you the same percentage used as Windows reports.
Please note that this experiment doesn't help with performace. Over the time, if you create and delete files in Windows again, SCP will again show larger numbers. Also, this experiment would not work if encryption/Bitlocker is activated.
I suspect that the storage problems you experienced earlier had to do with the bug that your storage grows over your plan limit. This should be fixed, at least on my side it did not occur anymore within the last 24 hours.
I will try to use the server as I normal, and see if it will happen again.
Yes, my servers also have that message in the SCP when they are stopped. While running with Debian 12 it looks like that:
# systemctl status qemu-guest-agent.service
● qemu-guest-agent.service - QEMU Guest Agent
Loaded: loaded (/lib/systemd/system/qemu-guest-agent.service; static)
Active: active (running) since Wed 2025-08-06 18:33:22 CEST; 1 week 5 days ago
Main PID: xxx (qemu-ga)
Tasks: 2 (limit: 19046)
Memory: 2.4M
CPU: 13ms
CGroup: /system.slice/qemu-guest-agent.service
└─xxx /usr/sbin/qemu-ga
Aug 06 18:33:22 myserver.mydomain.tld systemd[1]: Started qemu-guest-agent.service - QEMU Guest Agent.
Alles anzeigen
Yes, my servers also have that message in the SCP when they are stopped. While running with Debian 12 it looks like that:
Code Alles anzeigen# systemctl status qemu-guest-agent.service ● qemu-guest-agent.service - QEMU Guest Agent Loaded: loaded (/lib/systemd/system/qemu-guest-agent.service; static) Active: active (running) since Wed 2025-08-06 18:33:22 CEST; 1 week 5 days ago Main PID: xxx (qemu-ga) Tasks: 2 (limit: 19046) Memory: 2.4M CPU: 13ms CGroup: /system.slice/qemu-guest-agent.service └─xxx /usr/sbin/qemu-ga Aug 06 18:33:22 myserver.mydomain.tld systemd[1]: Started qemu-guest-agent.service - QEMU Guest Agent.
Thanks! That means the 'qemu-guest-agent' part is not a concern. I will focus on the storage between the Netcup panel and the Windows Server.