Recent content by mfed

public_network is what is used by the Ceph clients to connect to the Ceph data. I believe you need to move your public_network to your 10Gb/s subnet

I also use Mikrotiks for my network, but the MLAG was unreliable at that time I tested a few years ago, so I used l3hw capabilities of the switches and created a redundant routed ECMP configuration. Works great for me...

Hi @ralphte I am not a Ceph expert, but to me it looks as if you removed and re-added a host with the existing OSDs to the cluster. If you do that the host might change its id, that will make the crush algorithm to produce a different desired placement for a lot of pgs. I see you don't have any...

Hi @jpv31 Not sure about the GUI, but you should be able to move all VMs from a bad pve1 server to a good pve2 server from CLI using a command like that 'mv /etc/pve/nodes/pve1/qemu-server/*.conf /etc/pve/nodes/pve2/qemu-server/'. Then you should be able to start them.. The etc/pve is a...

Yes, that's when you would need the 'pvecm expected 1' for the recovery. Better IMHO than using 'pvecm expected 1' permanently...

In theory, you should be able to modify corosync configuration (/etc/pve/corosync.conf) and assign to the node that you want to keep enough of votes so that the single node will be the quorum with no regards to the other ones (e.g. set quorum_votes to 4, the other 3 nodes with quorum_votes of 1...

Not a huge Ceph expert, just a long time user in my home lab cluster, you just pick up various small things as you go... I setup my cluster when there were no autoscale functionality, there used to be different calculators, that's where I know that it is supposed to be a few dozens pgs per OSD...

Maybe it just not able to finish all the scrubbing operations for all your pgs in two weeks? I had it when scrubbing stopped due to some OSD issues, and it took maybe a week or more after the issue fixed for the pgs to get all scrubbed... Also, to me 97pgs for 64 osd seem too low... They say...

Yep. I run my cluster in this configuration for more than 7 years and still waiting for something to happen... It's a lab anyway, I actually wanted to see what could go wrong as a learning exercise, but it just works...

I have a 3-node proxmox/ceph cluster with consumer grade NVME SSD and it works fine and I use a dozen or so of different VMs. Just checked at my 1TB Intel 660p and Crucial P1 that I started to use in 2019, one of them has 108 TB written, the other 126TB. Basically that is less that 2/3 of their...

@complexplaster27 For each OSD you configure the address to use, something like below (I use different cluster and public subnets, but you can just use the same address for both). Then you restart that OSD (osd.8) to start using the new IP address. Check that everything works, the cluster still...

I think you should give it some time for the deep scrubbing to finish and fix the inconsistencies.

@complexplaster27 No, it's not related to proxmox network. When you restart an OSD configured to use the new subnet you will need it to be able to communicate to the OSDs that are still on the old subnet. Same with the MONs, when you re-create the new one still need to communicate to the old ones...

Hi @complexplaster27 I responded to similar questions on the forum. I believe you will have to ensure there is a routing between your old and new subnets for the duration of the transition. You are correct, you just modify those parameters appropriately as you go (I believe there is also...

I would say the very first step would be to configure your new network and ensure hosts can talk to each other on this network. I believe that all can be done from GUI and not related to ceph at all. You will also need to enable routing between the old subnet and the new subnet. During the...