Esxi free is dead. Alternative?
- Thread starter Nemesis_001
- Start date
Notice: Page may contain affiliate links for which we may earn a small commission through services like Amazon Affiliates or Skimlinks.
Proxmox is amazing!!
It has incredible features including replication, migration, snap shots, lxc containers, linked clones, etc.
You can build a 2 node cluster and the sky is the limit on what you can configure.
You will never miss ESXi ever again, and if you really need to you can just run ESXi nested as a guest vm
It has incredible features including replication, migration, snap shots, lxc containers, linked clones, etc.
You can build a 2 node cluster and the sky is the limit on what you can configure.
You will never miss ESXi ever again, and if you really need to you can just run ESXi nested as a guest vm
SmartOS would be as compact and efficient as ESXi with more VM options (Bhyve, KVM, Linux LX container/Docker, Solaris zones) and has native ZFS support. But I suppose Proxmox is the future.
If you want to seperate VM services and storage services for best availability or short recovery, a full featured storage VM under Proxmox remains an option, be it based on BSD, Linux or OmniOS/Solaris. A minimalistic resource efficient OmniOS with its kernelbased SMB server (nfs4 ACL, Windows SID as file attribute, local SMB groups) is an alternative to SAMBA based solutions there too.
I have not found the time to evaluate OmniOS Unix under Linux Proxmox.
The newest OmniOS has included virtio-scsi driver. Not sure if this is important.
If you have tried newest OmniOS under Proxmox, please report results.
If you want to seperate VM services and storage services for best availability or short recovery, a full featured storage VM under Proxmox remains an option, be it based on BSD, Linux or OmniOS/Solaris. A minimalistic resource efficient OmniOS with its kernelbased SMB server (nfs4 ACL, Windows SID as file attribute, local SMB groups) is an alternative to SAMBA based solutions there too.
I have not found the time to evaluate OmniOS Unix under Linux Proxmox.
The newest OmniOS has included virtio-scsi driver. Not sure if this is important.
If you have tried newest OmniOS under Proxmox, please report results.
Last edited:
I am trying to run it under XPC-NG (
) which is actually XEN, and I like it much better the PROXMOX. But after install of OmniOS I get the following error - screenshot below.
unable to configure /xpvd nexus
Seems to be some problem with apix module.
serverfault.com
The interesting thing is that there is an OmniOS image for AMAZON AWS - AWS EC2 Image which runs on XEN (or did they go to KVM?) so it should actually run? How can I get it to run... it seems very close.
I tried with apix_enable=0 in /etc/system and all without success.
I also tried to run Solaris 11.4 on XPC-NG, it crashes in the setup, before I can do anything.
unable to configure /xpvd nexus
Seems to be some problem with apix module.
Can't install Solaris 10 on XenServer 6.5 VM
I have XenServer 6.5 installed on a server, and i have been trying to install Solaris 10 on a VM, it just get into the boot screen of the installation ISO, i proceed with ENTER and then this message
The interesting thing is that there is an OmniOS image for AMAZON AWS - AWS EC2 Image which runs on XEN (or did they go to KVM?) so it should actually run? How can I get it to run... it seems very close.
I tried with apix_enable=0 in /etc/system and all without success.
I also tried to run Solaris 11.4 on XPC-NG, it crashes in the setup, before I can do anything.
People are still using VMWare???
KVM based
Proxmox
Nutanix (there is a free homelab capable version)
RedHat Hypervisor
XENish
OpenStack
XCP-NG
honestly ProxMox is the best all around platform. All the others have uses, and fit some needs better, but the homelab, small to medium business, and people getting out from under VMScare, ProxMox is it.
KVM based
Proxmox
Nutanix (there is a free homelab capable version)
RedHat Hypervisor
XENish
OpenStack
XCP-NG
honestly ProxMox is the best all around platform. All the others have uses, and fit some needs better, but the homelab, small to medium business, and people getting out from under VMScare, ProxMox is it.
I was afraid this would happen, so I've been deploying TrueNAS Scale to new small business customers as an option were HA isn't that important.
I've been playing around with a Proxmox standalone host for a while and really liked it. I'll be trying it out as a cluster soon. But I also want to take a look at XCP-NG.
It's very likely I'll just move to Proxmox and forget about renewing my vmug subscription.
I've been playing around with a Proxmox standalone host for a while and really liked it. I'll be trying it out as a cluster soon. But I also want to take a look at XCP-NG.
It's very likely I'll just move to Proxmox and forget about renewing my vmug subscription.
thats ProxMox.If only XCP-NG would give you native management on the nodes, instead of having to run a Xen Orchestra VM just for management
literally the fact that Xen Orchestra is separate from the host is its single selling point in comparison to Proxmox. you can manage a bunch of sites from one server running Orchestra somewhere. if you want a hypervisor that includes the best management interface you can have and you want it to reside on the host, that is called ProxMox.
Nutanix actually kind of splits the difference. but the free one is behind the paid one, and the free one does not have the same size community as ProxMox.
It does. Its called XO Lite and you have native management of the node itself, currently start/stop console, some statistics, but still missing some other features, but it is being developed to be available soon with full feature set.If only XCP-NG would give you native management on the nodes, instead of having to run a Xen Orchestra VM just for management
I love the nutanix platform, and their CE (free) edition is exceptionally generous (must continually be updated to the latest release and up to four nodes only - which isn't a problem for most home users), but do keep in mind that it's pretty resource heavy and expect to dedicate (as in entirely) at least 4 to 8 cores on each node and 24 to 32gb ram. Again - these are 100% dedicated to nutanix controller (CVM) VMs. You can run it with fewer resources at the expense of major performance losses or loss of functionality, such as deduplication and erasure coding. These requirements make running the Nutanix cluster on most consumer-grade equipment is a pain.
Last edited:
Thanks for all the feedback.
I suppose the ground isn't burning, and existing esxi still runs fine and will continue to do so for years.
But I suppose it's time to start messing around with some of the other solutions mentioned.
Going to need another piece of hardware for that.
I suppose the ground isn't burning, and existing esxi still runs fine and will continue to do so for years.
But I suppose it's time to start messing around with some of the other solutions mentioned.
Going to need another piece of hardware for that.
Would be fine if those who use ProxMox or XCP report experiences and tipps with a Unix storage VM (OmniOS/ Solaris) or Windows (ZoW)
- Installation (problems/ workarounds)
- Tuning (nic, memory assignment)
- Consistent ZFS snaps (like ESXi quiesce snaps)
- Backup/ fast disaster recovery (configured storage VM best via snap/zfs send methods and bootenvironments)
- Installation (problems/ workarounds)
- Tuning (nic, memory assignment)
- Consistent ZFS snaps (like ESXi quiesce snaps)
- Backup/ fast disaster recovery (configured storage VM best via snap/zfs send methods and bootenvironments)
Last edited:
I think we discussed that a little in hardwareluxx forum some months ago. My understanding (and I am not an expert neither with proxmox nor with EXSi) is that you won't get snapshots that are perfectly consistent with the state of the guest OS by simply doing a ZFS snapshot of the running guest. You would either need to shutdown the OS or also copy the guest memory. Proxomox currently only supports that for snapshots (which are not integrated with PBS) but not for backups made in snapshot mode. So I think you can do safe snapshots with proxmox but it's somehow not what you want to backup your guests :/
My journey has been moving from proxmox to bare metal + docker. I run all my services (more or less) on the same host with multiple docker containers. I utilize the macvlan driver frequently to assign unique and different ip:s for each container.
What i miss is some control of the hardware that you get in proxmox. Theoretically i think you gain more efficiency using vms/containers + docker. On the other hand it is easier to maintain, update and backup the docker containers.
What i miss is some control of the hardware that you get in proxmox. Theoretically i think you gain more efficiency using vms/containers + docker. On the other hand it is easier to maintain, update and backup the docker containers.
ESXi quiesce snaps at least guarantee that no writes to the guest filesystem are done during a snap process via vmware tools. This cannot guarantee that all dependend transactions are completed in a snap like with the hot memory option in ESXi but the chance of a corrupted guest filesystem in a ZFS snap with partly written atomic writes (ex data+metadata) is minimized.
As far as I know, filesystem freeze options are available in Proxmox (qemu guest agent or other method?)
If you need 100% guarantee that a ZFS snap is consistent, the VM must be off.
Only if a ZFS snap can quite guarentee that the VM filesystem state is consistent, you can ZFS snaps and replication for backup or version recovery/rollback with zero wait time (much faster and more space efficient than backup/restore methods based on a file copy)
As far as I know, filesystem freeze options are available in Proxmox (qemu guest agent or other method?)
If you need 100% guarantee that a ZFS snap is consistent, the VM must be off.
Only if a ZFS snap can quite guarentee that the VM filesystem state is consistent, you can ZFS snaps and replication for backup or version recovery/rollback with zero wait time (much faster and more space efficient than backup/restore methods based on a file copy)
Yes I also understand Proxmox backups in snapshot mode do what you call a "FS freeze".
If you do regular snapshots (not backups in snapshot mode) using Proxmox for VMs with their disks stored in ZFS, Proxmox uses ZFS snapshots, but I don't know where the memory dump is stored. Backup of those snapshots via send/receive would be your job like finding out about the mem dumps.
If you do regular snapshots (not backups in snapshot mode) using Proxmox for VMs with their disks stored in ZFS, Proxmox uses ZFS snapshots, but I don't know where the memory dump is stored. Backup of those snapshots via send/receive would be your job like finding out about the mem dumps.
Just to say, we have been using ZFS snaps for a decade, and when tested they always worked, yes, it was is if the power was unplugged, but it did load the VM and run from the backup.
btw, somehow I dont like proxmox. Seems like a house of cards, thrown together with a hope that it will work.
btw, somehow I dont like proxmox. Seems like a house of cards, thrown together with a hope that it will work.
mine regularly runs 400 days before i do any maintenance on it whatsoever. i JUST on Febuary 10th went from ProxMox 6.4 to 7.4. and upgraded RAM, moved to LSI 9400 series cards passed through to TrueNAS CORE.
My day job is in I.T. and accross the board i have been VMWare free for over 5 years now. ProxMox at home for probably 10 years, Nutanix at work.
Proxmox, TrueNAS CORE with VIRTIO and Guest Additions. (yes i just rebooted it, like i said, 400 days and then i go through and do all my updates across the board.) this is a TrueNAS core nightly, yes it is my live in use home storage, yes it is still named FreeNAS)
on the RAM USAGE:
i disable balooning and lock a large amount of RAM to the ZFS NAS
Last edited:
This is just making it worse with another layer that is not perfectly synced. Snapshotting is not the problem. Guaranteeing that the data in the snapshot is perfectly consistent is a problem and you can not solve that by changing the backend. ZFS supports perfect synchronized writes. The FS won't be corrupt in itself. It's the software that might have data in flight because it is not programmed for situations that are somehow like pulling the plug without knowing when.