Search results

This sounds like a duplicate IP address problem to me. Try to compare the MAC addresses in your workstations ARP table when this problem occurs. My guess would be, that the MAC address changes when you have these short outages.

Hallo, ja das ist prinzipiell möglich, das PVE verwendet "nur" ordinäre LVM disks dafür. Online gibt es genug tutorials zum Vergrößern on LVs – müsste aber sowas wie lvextend -L +1G --resizefs /dev/<LV> sein. Wenn du Glück hast, kannst du das sogar ohne einen Neustart etc. machen. Ich hoffe...

Thanks for your detailed answer :) Is it possible that you have a duplicate ip address in your network? Check if any other device has the same ip address as one of your nodes.

These state changes are completely normal as far as the last one is "forwarding state". That means, the VM is able to communicate via the VPN for a few seconds and then it stops working? Would be very strange... I see no reason for the VPN to stop working. Does a bare metal client behave the same?

From what I can tell, the VPN endpoints are able to communicate with each other and establish a connection :) After 30 seconds the service terminates (I assume you stopped it?) What is the problem now? The VM is not able to communicate via the VPN? Are there any other suspicious system logs...

This is a very broad problem description... There are many reasons that could prevent "VPNs" to work. What type of VPN are you using? (IPsec, OpenVPN, ...) Is there L3 connectivity between the two VPN endpoints? (e.g. can they ping each other) What's in the logs of both endpoints regarding...

Sounds like a misconfigured bond. - How is the other bond endpoint configured? - Are VMs on the same host/vmbr able to talk to each other normally?

Can you only not connect to the GUI, or is the server entirely unreachable? Why would you obtain an IP address with DHCP, when there is a statically configured address? Is the statically configured address and gateway correct? Check before and after the DHCP request (after a reboot): ip a and ip...

There should be no problem switching to PermitRootLogin prohibit-password or similar. Note: Doing so prohibits new cluster joins, for these you would have to temporarily allow root login via password. After the join this will not be needed.

I think, you mixed up your net0/net1 config: Currently net0 is attached to vmbr10 and would correspond to ens18 (which should be vmbr0, I think?) To get the names of the interfaces use ip a (short for ip address) Furthermore, there are plenty of tutorials online for configuring IP addresses in...

This looks not too far of :) Now you have to add to each VM which needs this new subnet a new NIC (tab "Hardware" on each VM) with vmbr10. Don't forget to configure the ip addresses for this new network in the VMs. Hint: It would have been possibly easier to add the new vmbr10 via the GUI than...

What is your setup then? For me works this: - SPICE (qxl) graphical device in VM - SPICE or noVNC console - to a Windows guest with SPICE drivers installed (See docs)

Try enabling "Use tablet for pointer" under the "Options" tab of the VM. This small tweak enables absolute mouse tracking. See KVM docs (and search for "tablet") for more info :)

Es wäre sicher nicht schlecht an das Journal zu kommen: Wenn du das Journal auf persistent schaltest (dafür gibt es genug Tutorials online) kannst du mit z.B. journalctl -b -1 das Journal vom letzten Boot ansehen. Nachdem du das gemacht hast sollte beim nächsten Crash zumindest das Journal mehr...

This service probably relies on NTP (Network Time Protocol) which uses UDP port 123. Deactivating this service is a rather bad idea, because many applications require the participants to have the (nearly) exact same time. Is this port blocked by a firewall or something similar?

Thank you for your feedback. A fix has already been applied and will be available with the next update :)

Thank you for your feedback. A fix has already been applied and will be available with the next update :)

So wie sich das anhört ist mit Sicherheit nicht der Guest-Agent das Problem. Es ist wesentlich wahrscheinlicher, dass die VM (aus irgendeinem Grund) abstürzt und dabei halt auch (natürlich) den Guest-Agent mitnimmt. - Was genau ist das für ein Stick? - Wie hast du ihn durchgereicht? - Bitte...

Yes, this is possible :) See https://pve.proxmox.com/wiki/VM_Templates_and_Clones This means, you have to install the OS on at least one VM once, convert it to a template and use this as base for new VMs

Das kommt darauf an, wie flexibel deine Lösung im Endeffekt sein soll – in Linux gibt's viele Möglichkeiten. Jedenfalls brauchst du eine neue vmbr (ob Linux oder OVS ist egal). Diese vmbr muss auf "VLAN aware" gesetzt sein. Dann kannst du en1 als Port auf diese Bridge hinzufügen. Bei den VMs...