VM replication Job doesn't create VM on second node

[user8923428934]

Hi,

We have a 2 node proxmox cluster with the identical zfs pool on both. when I tryx to configure replication for a VM it starts the Job and finishes the job with success. (The Vdisk gets created on the second datastore) but The VM is not created on the second node.

Can anyone help me with this?

See screenshots

2 replication jobs active opn node 1


However no VMs created with these IDs on node 2

But both vdisks on target storage on node2
Thanks

 

[sterzy]

Hi,

replication just replicates the storage, but you will still have only one VM. Migrating it to the other node should be faster though since the storage is already present through the replication process. For this reason the corresponding chapter in the manual is also entitled "Storage Replication" [1].

[1]: https://pve.proxmox.com/pve-docs/pve-admin-guide.html#chapter_pvesr

 

[user8923428934]

ok thanks for the info . but isn't this meant to be used in a node failure to start the vm on a new node? beacuse if I only havet the disk on the traget node how do I then start it (ssuming the 1 node failed)

 

[sterzy]

Ok so the idea behind storage replication is that you have redundancy for VMs that use local storage. The storage of the VM will get replicated to another node, but you won't have two VMs running. This is necessary because two VMs modifying the same (virtual) disk at the same time would cause inconsistencies and instability.

To automatically re-start the VM on a new node in the case of node failure you can turn on HA, but be aware that you will lose data that has not been synced yet. So if you replicate the VM's storage every 15 minutes, you might lose up to 15 minutes worth of data.

 

[Nixus]

Hi, i agree with user8923428934.

I thought it was more like (sorry for cussing) Hyper-V.
Please help us understand.

this scenario.
Node 1 active with runing vms.
Node 2 inactive with storage replicated from Node1, but no vm settings replicated.

Node 1 hangs, crashes or something similar.
How can i get the vms up and running on Node 2 if they arent even there to start?
Well except from the replicated storage.

I also enabled HA and did a test migration when both nodes where up.
went fine, but the vm migrated away to Node2 (per design i guess).
question remains though, if one crash or something else and we need to get te ones replicated to Node2 running.

Bare in mind that i am kinda new to Proxmox in cluster setup, onlu used single nodes until now.

All Best.
Nixus

 

[bbgeek17]

this scenario.
Node 1 active with runing vms.
Node 2 inactive with storage replicated from Node1, but no vm settings replicated.

Node 1 hangs, crashes or something similar.
How can i get the vms up and running on Node 2 if they arent even there to start?
Well except from the replicated storage.

If node1 and node2 in above scenario are part of the _same_ cluster, then they have a shared configuration filesystem in common PMXCFS (https://pve.proxmox.com/wiki/Proxmox_Cluster_File_System_(pmxcfs)). That means the VM configuration is known to both nodes simultaneously, but only one "owns"/runs the VM. So if that node fails, the other node can start the VM _if_ the storage is there, which you are achieving with replication.

---

Blockbridge : Ultra low latency all-NVME shared storage for Proxmox - https://www.blockbridge.com/proxmox

 

[Nixus]

Ahhh. thank you so much for clarifying this.

i wish you a great weekend.