Search results

https://pve.proxmox.com/wiki/Storage_Replication

waiting for them to repair ... patience ...

all the hosts (cluster and new one) can ping each other with their IP address and with their names ? have you made a first ssh connection from new host to each cluster host and from each cluster hosts to new one before starting to join the cluster ? Regards

:oops: When you reboot, does anything seems to show that the disk is still alive in the POST process ?

Proxmox is not a NFS Server. As Proxmox is Debian based, you can easily install a NFS Server in the Debian part (but in the shell, not with Proxmox GUI) and you can access to your NFS shares from all your host ... so your project is "technically" possible only with your 2 Proxmox hosts. But...

seems that your computer does not "seen" your disk any longer ... maybe your hd is out of order ... Try this command : lsblk | grep disk Is your disk alive ?

Your computer does not find any local media to boot on. So in it's Bios configuration (often) after trying to boot all local sources (hd, USB, DVD, etc ...), it tries to boot on the network using the BOOTP protocol. BOOTP tries to contact a server that will give it an IP address and a TFTP...

master can't be reached in SSH ?

iSCSI use blocks, it runs nearly like your computer talks to a local hard disk. iSCSI is like a local disk but connected thru the network but the network is only another cable like SATA or SAS connectors NFS is a network protocol used for sharing files with rights etc ... It runs on a server...

If you want to have the same storage place for all your hosts (for high availability or online migration), you must have a shared storage (https://pve.proxmox.com/wiki/Storage) : - a simple NAS running NFS share or SMB share (poor performance), this NAS is connected to the same physical network...

Lots of hard reset of the e1000 network card ... [615030.679729] e1000e 0000:00:19.0 eno1: Detected Hardware Unit Hang: TDH <1> TDT <e4> next_to_use <e4> next_to_clean <0>...

The 1 gbs last network card is dedicated to https/corosync.

Hello No the 10G bond is dedicated to SAN access. No other traffic on this bond. We can't afford a network congestion on the SAN network. Backup will go through the the 3 x 1 Gbs bond Regards

shutdown -r 0 (polite reboot) init 6 (polite reboot too, nearly the same as shutdown -r) if none of them runs shutdown -r -f 0 (agressive reboot) Regards

Did you open a ticket at OVH Support ? Maybe it's a netword card physical trouble or just the network cable that is not firmly engaged in its port ... Is there any info in syslog logs ? Regards

I'm thinking about a important point : is the proxmox firewall activated ?

type "more /etc/hostname" to know the name of your host check that you have in your /etc/hosts file things like 1.2.3.4 name_of_your_host 1.2.3.5 name_of_your_second_host etc ... or you can edit it thru the Proxmox GUI by selecting your server in the left column then system /...

Thats a problem ! https://vitux.com/linux-hosts-file/ As long as you cannot resolve by name in your cluster, you will have trouble ... Regards

Does the ping runs with the names of the servers, not the IPs (ping 3173716) ?

Ceph speed really depends on network bandwith : https://www.mellanox.com/related-docs/whitepapers/WP_Deploying_Ceph_over_High_Performance_Networks.pdf An interesting paper about Ceph (with Openstack but it doesnt matter in fact) http://www.cs.utsa.edu/~atc/pub/C51.pdf If you plan to agregate...