Suggestions on VPS/Cloud with HA/Failover setup

vpsgeekvpsgeek OG
edited December 2019 in General

I manage a VPS for a friend where currently 6 websites are hosted with DirectAdmin. Now he wants his websites to be on a HA/Failover setup where if vps goes down due to hardware issue with host or whatever reason, all websites should be up automatically on a reserve server with same specs. Out of the hosts I have had services with netcup offer it in form of failover IPs & probably Hetzner do as well in form of floating IPs.

Are those suitable for this kind of setup or do you recommend something else? This would be my first time working with such a setup
His budget is €50 per month which he can increase a bit if required to achieve this kind of setup on vps. This budget does not include cost of DirectAdmin & required specs are:

1 vCore
4 GB RAM
50 GB Disk
5 TB Traffic
Europe
Budget: €50

Recommend: SmallWeb|BuyVM|Linode|RamNode

Thanked by (2)uptime WebHostingDomain
Tagged:

Comments

  • Any cast DNS+Rsync+Database Replication across multiple VM. That would be your best bet if you want a true lowend style Failover Solution.

    Thanked by (2)vpsgeek uptime
  • FHRFHR Hosting ProviderOG

    If you really want to do proper active/active (master/master) replication, prepare for a journey through the dark, dark world full of pain.

    What you need is two servers, which are very close together (latency wise), but preferably not hosted with a single provider. Then you can play with stuff like DRBD, which can replicate block devices (volumes) across network - that's for your website files. Joined by MariaDB Galera Multi-master cluster to sync up your database, some sort of load-balancing / failover solution (think intelligent DNS, hosted by a 3rd party)… the list can go on.

    HOWEVER, seeing as he wants DirectAdmin prepare for even MORE pain. Since these control panels never support any sort of HA, you'll need to do it on a lower level, aka replicate the whole VM with the stuff, aka something which is hardly low end.

    Building a proper HA solution takes a ton of effort and money and it is usually not needed. I would reconsider if something like this is REALLY needed. Wouldn't using e.g. Hetzner Cloud, doing hourly snapshots and then having some script in cron, which says "if X is down for 3 minutes, restore snapshot Y to a new server and change DNS to point there and alert me" suffice?

    Thanked by (2)uptime vpsgeek

    SkylonHost.com High Bandwidth European Cloud KVM | AS202297

Sign In or Register to comment.