• jackoneill@lemmy.world
    link
    fedilink
    English
    arrow-up
    3
    ·
    1 year ago

    I have a rack in my garage, all my servers and my gaming rig are in it - fiber cables run through my attic to my office to connect the front end of my gaming rig (monitor and usb c hub with peripherals).

    My domain and all of my services are either VM’s on vmware or xcp-ng, or dockers on a VM. It’s all VM’s, except for the gaming rig and the Veeam backup server - they are bare metal.

    VMware VM’s are backed up with Veeam to a bare metal windows 2019 backup server (because its backup its the only bare metal server in the rack). First copy goes to internal RAID5 array. Second copy goes to ISCSI target that is netgear NAS in RAID5. Third copy goes to Wasabi.

    Most docker data is on the docker VM’s themselves, but if i need to mess with the data on the docker, that data is on a RAID5 Synology NAS. This gets backed up to Wasabi via duplicati. The system backup for the bare metal backup server also gets duplicati’ed to wasabi to make recovery from a site level disaster a little bit easier.

    Everything XCP-NG gets backed up first copy to an UnRaid NAS with 2 parity disks, then copied to Wasabi. I will eventually get a second local target - I have some promising candidates in the shop, I just need time to diagnose and repair them.

    Every Windows server and my gaming rig gets backed up direct to cloud with Redstor since I have extra space in the NFR bucket where I work and I manage that product. Gaming rig has no other backup so that’s nice (most game installs not backed up but everything else is). Other windows servers are backed up either via the Veeam chain or the XCP-NG chain, but it’s nice to have a secondary backup. I would do the Linux ones as well but Redstor sucks with Linux right now and it’s not worth the hassle.

    Plex media is on another UnRaid box with 2 parity drives, but no other backup. This content can all be re-downloaded via automated systems if the array ever fails. I have had the hardware that data is on fail several times but it’s always been recoverable, knock on wood.

    I am slowly working on phasing out VMware/Veeam in favor of XCP-NG, but it’s a long process.