Recent content by spirit

seem than freebsd have fixed the checksum bug with virtio-net recently (september 2015 / freebsd15) https://cgit.freebsd.org/src/commit/?id=5a8f33f52ae38bbf12f8f45b8e0f75166805c386 I don't known if it's already available in pfsense/opensense It's related to this bug...

Hi dominik, I'm going to test with a H200 soon. have asked to nvidia support for ai enterprise support (and also about MIG mode not working with windows vms)

I have build the 6.17 kernel with openzfs 2.4 , no more error. If somebody want it, I can share it. (it's unsigned, so not working with secure boot)

seem to be fixed here : https://github.com/openzfs/zfs/pull/17943 I don't known if it's already fixed in current proxmox kernel

a difference with snapshot mode is that the qemu agent is doing a fs-freeze/fs-thawt before the backup, so maybe you can try to disable this ? (they are an vm config option in vm guest agent config). also, not sure that it could help, if you using the agent agent , they are also an extra...

move the file (delete the vm config on source node) to be sure to not restart it again on old01 by mistake, when it's running on NEW01, or you'll break your vm filesystem

you can edit manually /etc/network/interfaces and add: iface <vnetname> address X.X.X.X/X I'll be merged with the sdn config on reload

the vm traffic is goinf through the vxlan (vxlan_remoteip ...), so it's use peers ip address. if you want to split management && vm traffic on 2 nic, you need to setup 2 differents ip address on each host on different subnet. (1 for management, 1 for peers). Also, don't need vmbr0 or vmbr1...

systemd v256 -> not on debian13, so not on proxmox

Running 5000vms with 100 nodes since 5years. (I have made the original evpn code for it)

About pdm, Disaster recovery is on the roadmap. Evpn central configuration is already available in 1.0. If you need other vpn mesh, they need to be added in pve sdn first

maybe your hardware raid have writeback cache ? what is your ssd model ? datacenter grade ssd with supercapacitor ? (zfs will run poorly on consumer ssd). 0% pressure is totally fine, it should be 0% to have good performance.

you shouldn't use consumer ssd like Samsung evo with zfs. zfs is doing synchronous write, and consumer ssd don't have a supercapacitor to keep sync writes in memory cache before writing the nand cell. (it's really something like 200~400iops on theses drive with 20000 iops with datacenter...

"I'm assuming defining static routes to get back to the VMs from external is perfectly okay, but we're going to see about getting bgp set up on our switches" If you want to reach the evpn network from outside, if need indeed routes to the evpn network. (so static routes, or bgp learned routes...

the "evpn controller", is deploying config for evpn+bgp (they are only a bgp daemon in frr, but different peers for bgp or evpn layer), with same ASN for both evpn/bgp peers. but if you need create custom bgp configuration (e-bgp, different ASN , different peers for bgp,...) you can add...