New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Comments
I paid 1 in gra earlier today
Interesting, I will try my luck with one KS-LE-E order.
@marc1976 I assume this was the 2x480 SSD version and you got extra 2 HDDs?
Yes
My GRA LE-LE was droolworthy - 1650v4, 64GB RAM, 2x 1.92TB NVME. Wish I had more invoices there.
I got 1 same config already and 6 more invoices
Lucky! Enjoy!
I got 5 le-b transfered from few ppl, let the idling begin, have use for 2 for now
I'm still looking for a KS-LE-B with NVMe. LIM//FRA
Me too impossibile to get one, all transfers to same few people. 😞
Just had to be fast and most is already gone by now. You probably have higher chances after a restock (unlikely) or BF.
++1
@plumberg sending you DM now!!
O boy..
Is it happening?
YABS. Immediately!
Thanks for all the tips regarding HA! Too much to quote everyone individually, but I really appreciate it.
I've also toyed around with the idea of setting up a 3-node Proxmox cluster, but I'm not too fond of managing CEPH. I've heard lots of horror stories about it, and it seems quite picky and specialized. In the end, this hobby should be fun, not just frustration.
However, what would be nice is to be able to simply move VMs between Proxmox nodes. Not HA, but just manually moving them to, for example, reshuffle resources. I heard this can be "easily" achieved by setting up shared storage—is that shared storage CEPH again? Perhaps this deserves its own thread.
Make it happen Daddy Rock! You don't want to end up on my Newsletter list which shows how much I love my servers.
FIVE!?
At least one of us gets a 🥖 delivery
Just use a sshfs mount, add it to your other nodes and you can use that to move vm snapshots around.
On the destination node, you just click restore.
Healthy amount.
Most of my beefy VMs / dedicated servers run proxmox. But i did not create a cluster.
Frequently i move my vms but i do not use live migration feature, i can live with a downtime.
I just take a backup using my proxmox backup server and restore it on another proxmox instance.
Good enough for my use case.
But from my understading there is no need for ultra low latency for most of the cluster features. Live migration does not require low latency.
Ceph would require low latency to perform well. I did not test, but i'm pretty sure you can use bunch of features across datacenters
You don't even need CEPH, Live Migration works even with zfs.
As long as it serves your needs. I for myself, like the SDN feature very much. It allows me to connect all VMs in a simulated L2 network using vxlan. This allows me to minimize the use of public ip's for VMs and secures most of my VMs by making me put them behind a OPNsense.
Local (per node) Zfs + proxmox live migration works quite well and is easy to set up.
you can also use replication jobs and then you will just have to transfer the delta later on.
But ceph is actually quite stable. I mean it complains a lot, both on my proxmox and on my k8s (separate hardware, ceph clusters, etc) but has yet to actually fail or eat my data
When you use the default 3 nodes with 2 replicas (so data will be on all 3 nodes) a lot has to go wrong before you lose data. Ceph is solid enough to force read only before that happens (i.e. you lose quorum)
Edit: I think there is also a datacenter manager in beta (?) that you can use to manage multiple pve without having them clustered.
If you don't mind providing some wisdom @NotFoundException .
I have a Proxmox VE with OPNSense running to manage NAT, reverse proxying etc.
I have just added a second node to create a cluster. Can I use the OPNSense on the original node to also manage VMs on the second node?
Yes you can, as long as you got a network across the hypervisors, so that the OPNsense can access VMs on the other host.
If I where you I would setup a wireguard connections between them and then setup a VXLAN zone with a MTU of 1384 and a VNet in the created zone. Now that VNet can be assinged as a network interface for all VMs in the cluster and they can communicate with eachother like they were in the same lan.
Looking for any KS-LE-B with NVMe in north america (BHS etc). DMs open
So are you speechless in front of your new DADDY-LE-B, or nothing happened?
Still waiting
Pants down already
I wonder if worth the cpu cycles to keep the scripts running? I don't remember if B-sides was restocked last year
Another day, no FRA delivery.
In 2 days, FRA will be declared as fallen.
I think my OVH VPS has been down for about 2 weeks now