[SOLVED] Replication VM error

[finux]

Hello everyone! I'm working on a project here using Proxmox, and the scenario is as follows:
I have 2 Proxmox servers, initially running a File Server (Debian) on one of them, and on the second one, I want to replicate this File Server every 30 minutes.
So, I configured a cluster with only 2 nodes, both with ZFS storage, and it's working perfectly. However, the issue arises when I try to enable VM replication for VM-100 (File Server), which is giving me an error related to space, even though I configured both nodes with the same storage size.
Note: I migrated VM-100 (File Server) from a physical machine, and I'm not sure if that has any implications, but I wanted to mention it. Here's the error message:

 

[Chris]

Hi,
can you please post the output of the following commands for both hosts:

zpool list -v
zfs list -o space

Also, on a side node: running a 2 node cluster without an external vote device is not recommended, as if one node fails, the other one will switch into a read-only state, please have a look at the docs for a detailed description https://pve.proxmox.com/pve-docs/pve-admin-guide.html#_quorum

 

[finux]

Hello Chris, so here is the output of the commands:

 

[Chris]

Okay, so you have space left but I assume your disk is thickly provisioned. If that is the case, you need the same amount of storage for creating the snapshot to be send over to the other host. Please find a the steps to check and a solution in this thread https://forum.proxmox.com/threads/error-in-zfs-replication.128966/

 

[finux]

Hi Chris, thanks for your help!
After disable refreservation, the replication worked. Thanks

 

[Chris]

Hi Chris, thanks for your help!
After disable refreservation, the replication worked. Thanks

Glad it works now, please mark the thread as solved for others to find solutions more easily, thx!

 

[finux]

Hello Chris, before wrapping up this topic, I'd like to clarify something: in the event of physical issues with one of the nodes and the system goes into read-only mode, is there a way for me to work around this? Can I at least use the DISK to create another virtual machine?

 

[Chris]

Hello Chris, before wrapping up this topic, I'd like to clarify something: in the event of physical issues with one of the nodes and the system goes into read-only mode, is there a way for me to work around this? Can I at least use the DISK to create another virtual machine?

No, in read only mode you are not able to edit any settings, there is a possiblility to temporarily overwrite this by setting the expected votes to 1 via pvecm expected 1, but this is intended for maintenance and disaster recovery only.

As stated in my initial post, the recommended procedure to avoid such a state is by adding an external vote device, which assures the second node to continue to work as intended. This can also be a low power device, but should have a low latency connection to the cluster nodes. Details on how to operate and setup this can be found in the docs https://pve.proxmox.com/pve-docs/pve-admin-guide.html#_corosync_external_vote_support

 

[finux]

Não, no modo somente leitura você não pode editar nenhuma configuração, existe a possibilidade de sobrescrever isso temporariamente definindo os votos esperados como 1 via pvecm esperado 1, mas isso é destinado a manutenção e desastre apenas recuperação.

Conforme afirmado em minha postagem inicial, o procedimento recomendado para evitar tal estado é adicionar um dispositivo de votação externo, que garante que o segundo nó continue a funcionar conforme pretendido. Este também pode ser um dispositivo de baixo consumo de energia, mas deve ter uma conexão de baixa latência com os nós do cluster. Detalhes sobre como operar e configurar isso podem ser encontrados nos documentos https://pve.proxmox.com/pve-docs/pve-admin-guide.html#_corosync_external_vote_suppor

Ok Chris, thanks again for your attention.