back to article Veeam debuts its Proxmox backup tool – and reveals outfit using it to quit VMware

Data protection software vendor Veeam has delivered its promised support for open source virtualization contender Proxmox. Proxmox support arrived last week in the form of Veeam Backup & Replication 12.2. As detailed in a What's New? [PDF] file, the data protection tool can now create immutable backups of VMs managed by …

  1. Khaptain Silver badge

    Proxmox is now undergoing tests in order to establish if it's ready or not as our next platform...

    We truly hope that it is Enterprise ready.. We have 3 years, due to recent purchases and bean counting necessities, before us so that's gives us plenty of time.

    1. Piro

      Plenty of others are in the same position.

      After 3 years, VMware could have a serious problem.

    2. wolfetone Silver badge

      All I will base the following on is 3 years of using it at home on two servers (running 24/7 with various different applications):

      Proxmox works. VMware are in trouble.

    3. damo2929

      been using it in production for nearly 3 years running our private cloud.

      7 Servers, 480 cores, 15TiB ram and 1.8 PiB of storage all powered by proxmox only.

      1. Chz

        We have zero doubts about the capabilities of ProxMox, it's the support we're not too sure on vs. Nutanix's offerings.

    4. ChipsforBreakfast

      Another vote for Proxmox

      I've had proxmox in production now for close to two years in three different deployments and I'm about to deploy the fourth.

      First was a 3 node cluster with 240 cores & 512Gb per host. Storage was via iSCSI to an all-flash Synology array over 2 x 10Gb ethernet links. Absolutely rock solid performance, stability and usability, the only annoyance was upgrades weren't as easy as they could have been but it was an annoyance rather than a problem.

      Second was a smaller 3 node cluster with 32 cores & 512Gb per host. That one used converged storage with SSD's in each server providing a total of about 60Tb of CEPH-backed storage. We segregated the storage network on 25Gb ports and kept the VM's on 10Gb ports. Also absolutely rock solid and very high performing. CEPH took a bit of getting used to but again, once done it just worked.

      Third came a four node cluster distributed across four different DC's. 32 cores/512Gb per host, 10Gb/sec interconnects with 25Gb/sec for storage and again, a CEPH array of NVME disks totalling 40Tb. That particular cluster survived a complete DC failure without missing a beat...

      Currently, I'm working on a new cluster which will have 3 hosts with 192 cores per host, 768Gb RAM and 60Tb of storage per host.

      All of those have used Proxmox's own backup solution and it's never been a problem. Fast, reliable backup, very efficient on disk storage (the dedupe is amazing) and good alerting for failed/completed backups. I target the backups to Synology NAS's and use Synology's replication, snapshot & immutability to replicate & protect the backups. As a last resort there is also an offsite replica that's managed via Proxmox Backup - we should never need it but it's nice to know it's there!

      All in, I'd highly recommend Proxmox. In licence costs alone it's saved me tens of thousands, never mind the lower cost of hardware achievable by using CEPH.

      VMware and Hyper-V should be worried. Very worried.

  2. Roopee Silver badge
    Thumb Up

    Proxmox Backup Server...

    The article forgot to mention it - Proxmox’s own tool. I use it for my Proxmox HomeLab, and it works, of course. It can even run as a VM itself, which is how I trialled it.

    Sadly it doesn’t do backing up of ESXi VMs and restoring them to PVE , which would have been REALLY useful when I switched from VMware to Proxmox! It also can’t backup my Windows office PC; I use the free version of Veeam for that.

    1. TReko Silver badge

      Re: Proxmox Backup Server...

      Yes, why pay for Veeam, when Proxmox backup server is included?

      https://www.proxmox.com/en/proxmox-backup-server/overview

      1. Paul S. Gazo

        Re: Proxmox Backup Server...

        With only having Proxmox's web page to go by...

        VBR has robust immutable support and Proxmox' page doesn't say anything about that. Also, VBR handles multiple source environments including physical workloads in the same management system, so you can back up physical servers plus VMs (from multiple hypervisors if need be) into the same storage system. Those are two off-the-top-of-my-head thoughts, but I'm sure there are more. VBR is a very rich and mature product and almost anything you might ask for it has... as long as you can cope with it being a Windows product itself, so you need at least a Win10 box/VM to run it. Yes, that's right... you can run it on a Win10 desktop. Oh, and VBR can integrate with Veeam for Azure, so if you happen to VMs or Azure Files or Azure Managed SQL instances, VBR can talk to all that as well.

  3. Tubz Silver badge

    I'm a Linux noob and even I managed to use an old Dell Precision T3420 + Proxmox+OPNSense+Adguard to replace my near EOL Asus router for a more capable and secure router and then just use the Asus router as an AP as it's wifi is more than enough. Next is to migrate my two Windows Severs boxes. Added advantage, I'm starting to learn Linux which I wouldn't have done if I stuck to Windows.

    1. katrinab Silver badge

      You might find you can do the Adguard stuff directly in OpnSense. It can certainly replace PiHole, I'm not sure if AdGuard has any additional features that can't be replicated in OpnSense.

      1. Tubz Silver badge

        Yes, Adguard added directly to OPNsense along with upnp, Unbound DNS and various other goodies, I like the fact I'm talking to the highest upstream secure DNS servers and not local ISP.

  4. harrys Bronze badge

    its really good in a sme setting where its my defacto goto virtualization platform for my sme customers

    though no need for veam... proxmox backup server itself is a really good backup system, it really dedupes the data down and backups are super fast

    i even run my pbs as a virtual proxmox vm instead of bare metal as it makes disaster recovery of the pbs itself a breeze

    its all made possible with the pbs datastore existing on a nfs share (not on the pbs itself) on an entry level synology nas with mirrored drives and rsync of the shared nfs folder elsewhere for that extra layer of backup... can also put a seconday synology nas at the directors house and keep them synced

    ps smaller sme customers are usually cash strapped... above scenario is perfect with 4/5 older servers and second hand synology's instead of having to get them to buy new, some times that the only way i can convince them to pay decently for my time!

    pps the easy peasy lemon squeezy immutability option of snapshots of shared folders on synology is a real sleep easy solution when thinking about ransome ware infecting your customers data and getting that dreaded call :)

    1. seven of five Silver badge

      This is an "AI" generated post? Or did someone drop his box of "tech gibberish"?

  5. J. Cook Silver badge
    Flame

    One of my projects this coming year is to get a sandbox set up in our test lab and do a 'back off' between proxmox and Nutanix as replacements for vCenter.

    due to how our company is placed, we can only get the Cloud Foundation product even though we don't use (nor do we plan to use) most of the features it has, and it's MUCH more spendy than what we were paying for support by a factor of ten. It's... insane. It also didn't help that broadcom was/is screwing over VARs with their asinine requirements programs.

    1. ChipsforBreakfast

      Proxmox will take a bit of getting used to but the effort is worth it. Migrating VM's in takes a bit of thought and planning but Veeam may well make that process considerably simpler (I've yet to try it).

      Cost wise it's next to unbeatable - you can choose the support level that best suits your appetite for risk, skillset & business need but even the top end is a fraction of the cost from the big names. Personally I tend to go for basic or standard as my team are generally able to handle most issues in-house and our setups are pretty static.

      Compare the cost for my dual cpu servers (96 core total):

      Proxmox Basic support - £680 / year

      Microsoft Hyper-v 2022 - £6348 with NO support, just the licence!

      I don't run vmware but from what I hear the cost difference now is even worse.

      Absolutely no contest from a financial perspective.

      I've yet to find anything I need to do with Proxmox that I haven't been able to do. Every use case is different but it's an impressive piece of software - it's become my go-to tool for virtualisation projects now.

      1. Charlie Clark Silver badge
        Thumb Up

        If they have to scale up to meet increased demand, you can expect support costs to rise somewhat. But, even, if they double and there's nothing to indicate they will, it'll still be a lot cheaper than the alternatives and businesses should be prepared to pay properly for good support.

      2. J. Cook Silver badge
        Boffin

        Vmware's cost... is rough.

        I asked for a budgetary quote from out reseller several months ago, and it was roughly five hundred US pesos per core for Cloud Foundation. For one year. We have a bit over one thousand cores in the enterprise. (Yeah. Ouch.)

        Hyper-V was on the long list of alternatives, but didn't make the short list for the bake-off because of the ridiculous amount of licensing needed. ( per core for the host for the OS, per core for the Hyper-V, and then per core for the guest OS, and if there's a SQL engine running on it, additional cost per core. It's complex enough that even microsoft's own sales reps have trouble with it. (We broke our reps trying to figure out licensing for thin clients many years ago, and that was an experience.)

        1. Charlie Clark Silver badge

          Ransomware?

          Hock Tan: Nice data centre you have there. Would be a pity if something happened to it, or, for example, it suddenly stopped working. But you can easily avoid that happening by sending $ 50 million, in Bitcoin, to my account…

          1. J. Cook Silver badge
            Devil

            Re: Ransomware?

            Pretty much- Multiply that by... well, most of the small to medium business organizations that went mostly virtual with VMWare, because at the time, it was a safe bet, and you have a LOT of organization's IT directors and CIOs deciding to take the hit and migrate to 'something else', either before their current support contract expires, or just letting it go out of support and taking that risk. (at least for the ones that purchased perpetual licenses like we did.)

            It's my opinion (and hot take) that Broadcom decided to alienate the majority of 'small fish' that provide rather a lot of their revenue to go after a small subset of whales that, while they pay a LOT of money, are less than the total amount of the rest of their customer base.

    2. Sudosu Bronze badge

      My lessons learned after running proxmox since version 3-ish to help you get started.

      Make sure you update the hosts file on each node with all other node, or have a super duper reliable DNS that won't die during an outage.

      Make sure you have a time source that all nodes can reach, and know to set the date from the CLI if you cannot access that time source during a recovery.

      I run NFS on hardware OmniOS with Napp-IT as the storage over a dedicated 10 gigabit fiber. This allows for compressed and thin provisioned storage and other ZFS benefits.

      Make sure you throttle your backup speeds across the nodes to prevent saturating the storage segment.

      As with many VM systems you may want to have guests frequently update their time from a trusted source to adjust for Any hardware induced drift.

      The quest agent helps interaction between guests and host...though I don't always bother to install it.

      If a node is removed, never use one with the same name again.

      I usually take down existing or find a spare couple nodes and reinstall from scratch and create a new cluster with new NFS live target when upgrading.

      Then I back up from the old node and restore to the new. Building the nodes is quick back up and restore can take a while...but it is a good test.

      Once it is stable I rebuild the remaining nodes and join them to the new cluster.

      Good luck in your testing!

      1. J. Cook Silver badge

        The storage I have for the bake-off is a NexSAN E18- a simple iSCSI block storage device. we also have an aged Nimble CS500 cluster (also using iSCSI) for migration testing that runs the rest of our test lab. The compute nodes are all Cisco UCS B and C series.

  6. Anonymous Coward
    Anonymous Coward

    Veeam: software for point and click admins

    We used Proxmox for over 10 years along with compiling ZFS for backups while point and click admins used expensive VMWare offerings and had to pay someone else to do their backups. Restored 36 locations from Cryptolocker in under 30 minutes total the one time a client got breached.

    Now that the point and click admins are using Proxmox it'll start being a bigger target. That's ok though. We're already using the next generation of virtualization and backup. There's no point and click interface. Sorry point and click admins...

  7. This post has been deleted by its author

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Other stories you might like