Hacker News Reader

Comment by digitalsushi | Hacker News Reader

I wonder how this will cut into the hobbiest/learn-at-home crowd with our half width, 1U free-tier-license ESXi servers we learn our skills for our day jobs on.

View on HN

I don't think this is a concern of theirs - the new owners don't appear to have any plans for it to be around long enough to need another generation of people to run it.

It's likely that for whatever products remain after this purge, there will be no more major releases, only security patches and ever-more-expensive extended support contracts.

I think people underestimate the long term impact of this kind of thing. Giving the people who use and maybe more importantly influence the purchasing of expensive commercial products the ability to easily play with them on their own is essentially free advertising. And taking that away after you've offered it is even worse since now those same people not only can't experience your product firsthand as easily, but they now actively dislike your company for screwing up their home lab.

Maybe the bet here is that VMWare is so entrenched or the time horizon of the people making the decisions is so short that it won't matter to them personally. But I bet long term it has significant impact and makes room for competitors with more foresight.

It's likely going to substantially grow the business of their competitors.

I'm already not going to renew my VMUG membership, and am planning to switch my VSphere cluster over to Proxmox.

Are you using live migration of running VMs, or wanting it to work in Proxmox?

Asking because in my testing of that in Proxmox 3 and 5 node clusters last month, the live migration would fail (aka just hang with no further progress) just over 20% of the time.

Haven't looked into it further (likely will though), but searching around online shows lots of other people having experienced the same issue. General sentiment seems to be "Proxmox is good for homelab and non-prod environments. But probably don't use it in production."

Check put XCP-NG if youre looking in this space. Its a mature xenserver fork with a much improved interface. Migrations work well, and it has a comparatively inexpensive vsan equivalent that works well if you prefer a hyperconverged route.

Thanks. I looked at it briefly, but the project doesn't seem to be headed in a good direction.

The main interface (Xenadmin) now has an EOL notice as their README: https://github.com/xcp-ng/xenadmin

The other alternative now seems to be a paid solution:

https://vates.tech/blog/introducing-vates-virtualization-man...

That doesn't seem like a good match for my needs. :(

I think you can use Xen Orchestra with xcp-ng, isn't it?

That's the second link I think. At least, when going to their web page, that 2nd link is what they point to.

Perhaps the performance of your storage or network isn't up to scratch. I have a 3 node cluster with ceph on ssds and a 10gb network and my live migrations worked perfectly fine.

This was all in a virtualised setup on my local workstation (Ryzen 5950X, 64GB ECC, local mirrored NVMe SSDs).

I've not seen flaky comms between virtual interfaces on this setup before, but you never know. ;)

Reply on news.ycombinator.com