New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Why do you use a dedicated server vs VPS from different providers
I'm re evaluating my tech stack after a dedicated server I had, hard drive failed and my backup plan wasn't designed well enough and I experienced catastrophic data loss, plus time and effort.
Every thing I did to recover the data just made it worse.
So now I'm thinking, do I even want a dedicated server.
Why do you choose dedicated over VPS?
Comments
Anything that can go wrong with a dedi can also go wrong with a VPS. You don't automatically get backups or anything like that with VPS, at least with the low end services that LET normally focuses on. You still have to do your own. You can get managed services ($$$) that include backups, but you can get that for either VPS or dedis.
Generally speaking -- its better to scale out than scale up. Meaning its better to have two smaller systems than one larger system. Its better to have four database servers than one large one. I personally wouldn't run anything production on a single system -- be it a VPS or physical server.
In this case -- its also prudent to keep redundant systems. I am not sure what your tech stack is, but a really solid pattern might be:
Assuming you have a small Flask application (Python)/Wordpress/or something inhouse developed:
I would also want any
codechecked into a version control system, and any large artifacts/databases backed up to an object store like S3.A single server isn't enough to have a highly available project -- even if its the best of breed with all the enterprise options at a major datacenter. If you want to describe your tech stack a bit more -- im sure folks here could help make suggestions. Most VPS hosts here would also be more than happy to discuss options that would help you build a more stable infrastructure as well.
I agree with you both. Trying to figure out cost of time and effort and time actually producing instead of fixing all the things
The time to set this up is technical debt.. If you were setting this up net-new, I would highly recommend you take the time to consider either:
You should be able to setup your stack on any host that offers a reasonably similar version of the OS you target within 30 minutes. More than happy to chat with you for a few minutes on Discord/Skype to review your environment (pm me). Ideally you could work with someone to get some configuration management setup (Ansible) and tested for you, and have a solid backup plan.
If nothing else; give you a path to consider moving forward.
For dedicated, its good for when application requirements increase over time, or above documentation (I'm looking at you Atlassian!), as you'll have more total RAM or have the option of adding more RAM easily without needing to move to another server. None of the dozen 2GB VPS' I have would be adequate and would require moving to another VPS server with 3 or 4GB allocated (perhaps more RAM can be added for additional monthly fee, I just don't roll that way, I'm promo pricing all the way).
If a server had tons of hand customization and tuning, I'd want to have an image based backup of the dedicated and know I'll be back up in hours (the speed of data transfer). If it was just applications, config and data, I'd want VPS's with simple backups that can be backed up and restored with scripts.
My home Windows desktop machine is image backed up. It would take days/weeks/months to reproduce my tuned setup over years of undocumented changes. On my laptop, I just have some common applications (ie, one stop at ninite.com) and some data I can save on dropbox. No image backup needed.
For me, with how cheap dedicated servers are nowadays they are just a way to take all the responsibility in my own hands, configure redundancy the way i want it, configure backups the way i want it etc. Because i do not want to rely on some vps provider doing this, especially with how price-limited they are nowadays.
VM-s (or containers) are good, but it is really nice to have full access to hypervisor, underlying storage etc to be able to make proper backups in a way that allows fast and easy recovery. Also, this way i can manage risks the way i want, like some things need multiple redundant backups and some HA setup, others can live on raid5 with simple weekly backup just fine.
And if i do not want this responsibility in some case for some reason - managed services are the way to go, using good well-known provider and minimum needed service, like shared hosting when possible, or vps if needed.
So IMO if you are interested in becoming better at it you should learn from your mistakes, setup things better this time and continue using what you were. If you just want things to work, however, simply get managed service.
Also, i had near-miss recently too, i got lucky and i was risking only not-really-important data, but i did not like it anyway. HDD failed in home pc (non-raid), then HDD failed on server when i was restoring backup (soft R5) but luckily 3 other drives survived both restore and rebuild. I had redundant backup of really important stuff, but reinstalling/reconfiguring OS would have been massive PITA, so i am now in process of searching for a place to store one more copy of those backups so that i am not relying on such improbable luck anymore...