Search results

I have used it on just about all my PVE & PBS going to PVE 8.2 & PBS 3.2 as just about all of them had the Intel interfaces that have that virtual function causing the "change" that the System(&@#$(%&@#$%&(D implements I have loads of )&#@$(%)&@#$)%(& words for SystemD, as the correct way for...

Busy FAFOing with this in a totally meshed setup, and the use case is to have Ceph on that SDN created VxLan/EVPN vnets of it's own. Reason to NOT have it on the "public" network, but also to allow VMs (pointing fingers to K8s) that wants direct Ceph storage access on the access network, and in...

Thank you, for explaining the reasoning! Hmmm... interesting statement, given that there are still updates on OpenVSwitch/OVN in the last couple of months with 3.3.0s release. Which also asks for something else, to understand direction and this comparison of DPDK vs VPP in OVN, is there...

Is there technical reasons it can't work with OVS, or just a matter of ProxMox made a decision it'll be more cost effective to not also bother with OpenVSwitch for SDN? Hmm.. would like to have a look at that, as I have the need/use of a config server ala AWS, with some IP that is...

Ticket number/URL? Have similar requests - don't want to duplicate :)

RTFM the 8.2 Release notes, and then use this https://github.com/D4M4EVER/Proxmox_Preserve_Network_Names from https://www.reddit.com/r/Proxmox/comments/1cd1quu/proxmox_ve_82_proxmox_backup_server_32_breaking/

Yeah, why read Release notes, the devs don't know what they are talking about and we SysAdmins just rtFm when it's an Emergency :shrug:

How many ProxMox nodes you have in the cluster doing this? if you don't have 3+, then how does this give you the redundancy?

Yes 100% correct! But having only a single switch and LACP/etc. stil is a SPOF on the switch, thus the idea is/was to have the direct linked interfaces too, as we are moving towards a leaf-spline/"CLOS". Even more "fun" is when that swith ports dis connect, you "lose" (interface down) the IP...

yeah, there are a coupld of those that only matches IPv4 (assumptions?) in stead of both IPv6/IPv4, but the "fun" is that (ie. OSPF/BGP IDs) are 32bit, and thus "assumed"/simplified as 4x octets, and then "typically" using the loopback/highest IP as the ID (assumptions) so ... yeah, IPv6...

True, I'm looking at it from the perspective of a switch/network link failure , so that is an extra link... Og, OpenFabric routing protocol doesn't work with more than 2 nodes on the same link - design criteria

Question: Have you tested/used OpenFabric with more than 2 nodes on the same interface (ie. a vlan/ethernet switch)? ran into a problem which I suspect makes it outside of OpenFabrics design criteria, so as I've implemented your stuff, it worked, but then found troubles on the links where there...

Instead of OpenFabric, have you tried this with ISIS?

Something I am missing, is a pre & post FRR config (perhaps templates that have values/numbers filled in from the GUI/SDN configuration like the ISIS process/router name, ASN numbers etc.). Loading those into/via vtysh/yang/etc., would ease custom and north-south configurations and...

Good day, I have new set of clusters to be deployed in environments where I do have needs for SDN (vxlan-evpn) and loopback routes for CEPH while preparing for IPv6 only networks. Though I did got it going at one stage (IPv4 only), there are still missing pieces/processes that I'm stumbling on...

oh, and there are another "bug" in the parsing/do of the Network ID of the ISIS plugin that fails FRR parsing as the GUI/plugin doesn't parse that before hand and doesn't report that error in the task output ;(

@spirit now for related IPv6 (and perhaps openfabric?) parsing of the interfaces too? Perhaps enhancing the interface to know/enable IPv6? https://bugzilla.proxmox.com/show_bug.cgi?id=5343

vlans is separate broadcast domains, in my case I have both a secondary gateway (LTE) that issues DHCP only, as well as a dumb/unmanaged switch on on the same physical interface. So I use it as a secondary/tertiary corosync interface and management, and thus need it stable/fixed, but need the...

and right there is the problem... hacking and having the ability to have run away processes ;(

problem isn't syncing to a new machine, just stating the obvious that I can today, and especially with the new machine, run multiple vzdumps to PBS server(s) without killing the hypervisor, thus the single global lock for vzdump is a tad..... overkill and causing other problems with the...