Search results

I'll try to add support in the gui soon

The sdn configuration shouldn't overrwrite the ip address. the config is generated in /etc/network/interfaces.d/sdn , and merged (not overwritten) with interfaces defined in /etc/network/interfaces . But, as with sdn , the vmbr0.101 is plugged to a vnet, you can't defined ip on vmbr0.101...

technically, this is the same. I don't see why vmbr0.101 syntax don't work for you. The sdn use vmbr0.101 syntax internally, that's why it's conflicting with your vlan101.

you can disable ssh encryption by editing /etc/pve/datacenter.cfg migration: insecure

yes, the "nat" checkbox on the subnet is creating the iptables rules.

a vnet is a bridge. as a vnet in a simple zone is internal only (not attached to a physical interface), you can reuse the same subnet present on another bridge. But don't use same gateway/bridge ip than your other nat bridge, or it'll break the routing.

You are luke to not have problem with this intel chipset before. They are a lof of reports since multiple years about tso/gso driver bug, doing nic hang see: https://forum.proxmox.com/threads/e1000-driver-hang.58284/page-8 Can you try: ethtool -K <interface> tso off gso off ?

The vlan1 on proxmox is the default vlan , for untagged traffic.

Thanks for the report . seem to be related. The ip range creation manual creation is expected currently.(I should add a note in doc). But the ip search in the range was working previously, maybe is it a regression in 4.X. I need to check that.

ok thanks ! (I don't have tried the 4.x yet, I'll try it this week)

short response: you need 3 datacenters for the monitors, and you can loose 1dc. (as you always need quorum for monitor) the osd can be located on 2 datacenters (replicat4 for example, 2 copies on each dc)

proxmox kernel is based on ubuntu (6.8 is from the ubuntu 24.04 lts). so yes it's fixed https://ubuntu.com/security/CVE-2024-1086

what is your ceph version ? ceph 17 (quincy) have a qos , to priorize client access vs repair ceph config set global osd_mclock_profile high_client_ops ceph config set global osd_mclock_profile balanced (default) ceph config set global osd_mclock_profile high_recovery_ops

I'll try to look at it next week. (I need to reinstall my netbox appliance). What is your netbox version ?

Yes, you need to use vxlan in this case. (or qinq (vlan in vlan) but I think that your hardware switch need to support it)

Each trunk port has the following defined VLANs: 100,110, 150, 160, 200, 250 do you mean your physical switch port ? Because you need to allow vlan30 too.

with the new proxmox nftables implementation ? if yes ,can you send the result of "nft list table bridge proxmox-firewall-guests" ?

just trolling ;) But Hetzner is the only hosting provider with such l2 crap network without filtering in known in the world. (personnaly, I'll use a routed setup if I was hosting at hetzner)

like iptables, statefull. (until you specify to disable conntrack, like "iptables -t raw -I PREROUTING -j NOTRACK" for example)

you can't use vxlan 2 for both vnet && in evpn zone at the same time. (I thinked we had a verification for this in the gui. Do you have done all the setup through the gui without editing manually the cfg files ?)