Recent content by rafafell

Hello everyone, This is a node recently added to the cluster. The pvestatd log message has been occurring since the node was started and repeats every 10 seconds. If I create from the `/run/pve` folder, the log message stops appearing. A strange thing about this node is that pve-cluster...

the /run/pve folder does not exist. when I create it, it is deleted when I restart

Hi everyone, I'm facing an issue on my Proxmox setup. The following error appears in the system logs: systemctl status pvestatd pvestatd[2105]: could not populate metric data cache: mkstemp "/run/pve/metrics.tmp_XXXXXX" failed: ENOENT: No such file or directory pveversion...

Hello @benoitc , I'm starting to implement CEPH here. Could you please report whether this arrangement of yours was successful. did it meet your needs?

@shanreich , Is it possible to backup a specific virtual disk or are you referring to the entire virtual machine?

Sorted out! Really were the switch ports that formed the vlan were configured with trunk/tagged for just one tag that they had not informed me. see also: https://forum.proxmox.com/threads/vlan-tagging.79830/

Hi @Stoiko Ivanov , my scenario is as follows: - 1 node with two NICs (one exclusively for public IPs - untagged/access; and another for private ips - tagged/trunk) - 5 node with one NIC (tagged/trunk) ** I don't have direct access to the switch that connects these machines (I have to ask...

Hi, When I put the ip of the node directly in the physical interface or in the vmbrX everything works fine, but when I pass this ip to vmbr.100 (see below) I lose the connection with the other nodes in the cluster and when I restart the node (despite the web interface loading ) I can't log into...

Any tips or suggestions to address the above issue? I have no idea how to proceed

thanks @bobmc and @badji , I tried vlan and ovs (don't try sdn) but fail. I opened another post about it: https://forum.proxmox.com/threads/two-subnets-use-the-same-nic.109329/ because I think it escapes the initial question of this post. I saw your observations carefully, when moving forward...

Hi, I'm having trouble making two subnets use the same NIC and being able to pass to the cluster nodes scenery: - 1 node with two NICs (one exclusively for public IPs - untagged; and another for private ips - tagged) - 5 node with one NIC (tagged) ** I don't have direct access to the switch...

I have two LAN configured and apparently working fine. But I have the following problems: LAN1 (192.168.4.0/23) provides correct IP for the vms of the local node, but I can't provide ips for vms of other nodes. I believe that if I linked this LAN to the physical interface of LAN2 it would solve...

thank you @bobmc and @badji , I will advance here and post here soon how they are going something else, in relation to network storage (ceph). I will have a limitation which is as follows: - three nodes are on one floor - three nodes are on another floor - I have a maximum of 6 ports to...

Thanks for the answer All nodes have four NICs, but I can only connect to this switch the NICs of the above drawing (I could at most put one more NIC on this switch). These nodes are on different floors (192.168.0.10-12 on one floor and 192.168.0.13-15 on another) interconnected by a switch...

the cluster is up and running (192.168.0.10-15), but I still haven't mounted the storage. I'm using local storage to think about the network issue first Storage must be zfs. And in a few months we'll switch to ceph with a dedicated switch