Hosting Service Backing up Websites and Disaster REcovery?
- Thread starter afadavit123
- Start date
- Status
PebbleHost
High performance, consistent uptime and fast support. Minecraft hosting that just works.
Hmmm.
That's kinda a tough one, there is no single answer to that question. It really depends on the type of service you are providing. For example, a dedicated server, personally I have a pretty hands-off approach allowing the person who is using the dedi to dictate their own level of data protection, all my servers come with at least 2 drives in tandem allowing for the client to either set up a RAID array or have the full capacity with either higher risk or offsite backups in place.
Web-Hosting:
With Web-Hosting I think it could go either way, if you are offering a budget or as cheap as it gets website hosting I can see a company not offering included backups, with that said I highly suggest you make VERY VERY Clear to anyone ordering/potential buyers in that process. Now if you are trying to offer an all-inclusive service which it sounds like you are looking to do the first thing you always want to do at a minimum is RAID, RAID RAID RAID! I highly suggest you do your own research on the subject but basically if you have a single drive failure, you're covered temporarily with zero downtime. The next step up from that would be offsite backups.
It kinda sounds like you are just focused on data and not losing data but remember entire servers can fail. Though it is not common, it does happen, redundancy can range from having a second ethernet cable plugged into a box to a full failover box both with dual PSU/CPU ETC setup.
Unfortunately, there is no single answer to your question, but as general as I can make it:
Redundancy is Key.
Rant over... it sounds like you are just starting off, I suggest looking for a reseller program where a lot of this is already tackled allowing you to focus on other issues. Francisco at BuyVM has been killing it for years and honestly, that's about as cheap as it gets: https://buyshared.net/cpanel-reseller-hosting/ in my opinion thats probably your best option to start. You can start on cheap shared hosting, work your way up through those then get to a VPS, and grow that larger and larger to the point where you have your own mini CDN network with Anycast.
If you are looking to build your own setup, play around in GCP, they have like 300 or something dollars in credit. Create some Location Mirrored buckets setup DirectAdmin and see what you can do.
-
Raaj
That's kinda a tough one, there is no single answer to that question. It really depends on the type of service you are providing. For example, a dedicated server, personally I have a pretty hands-off approach allowing the person who is using the dedi to dictate their own level of data protection, all my servers come with at least 2 drives in tandem allowing for the client to either set up a RAID array or have the full capacity with either higher risk or offsite backups in place.
Web-Hosting:
With Web-Hosting I think it could go either way, if you are offering a budget or as cheap as it gets website hosting I can see a company not offering included backups, with that said I highly suggest you make VERY VERY Clear to anyone ordering/potential buyers in that process. Now if you are trying to offer an all-inclusive service which it sounds like you are looking to do the first thing you always want to do at a minimum is RAID, RAID RAID RAID! I highly suggest you do your own research on the subject but basically if you have a single drive failure, you're covered temporarily with zero downtime. The next step up from that would be offsite backups.
It kinda sounds like you are just focused on data and not losing data but remember entire servers can fail. Though it is not common, it does happen, redundancy can range from having a second ethernet cable plugged into a box to a full failover box both with dual PSU/CPU ETC setup.
Unfortunately, there is no single answer to your question, but as general as I can make it:
Redundancy is Key.
Rant over... it sounds like you are just starting off, I suggest looking for a reseller program where a lot of this is already tackled allowing you to focus on other issues. Francisco at BuyVM has been killing it for years and honestly, that's about as cheap as it gets: https://buyshared.net/cpanel-reseller-hosting/ in my opinion thats probably your best option to start. You can start on cheap shared hosting, work your way up through those then get to a VPS, and grow that larger and larger to the point where you have your own mini CDN network with Anycast.
If you are looking to build your own setup, play around in GCP, they have like 300 or something dollars in credit. Create some Location Mirrored buckets setup DirectAdmin and see what you can do.
-
Raaj
Okay. thanks
Oops! My intention was not to make RAID sound bulletproof rather just dumb it down to a single line/statement, "but basically if you have a single drive failure, you're covered temporarily with zero downtime." I guess you could find that misleading, but I think most people who work with RAID Arrays will agree that in general, that is a true statement.
~80% Sucess rate? Jeez, where are you getting that from? RAID has gone a long way since the 90's. BQR did a study with RAID 5 on drives that have 1 failure/million hours and it had 99.72% Reliability, which is very very far from 80%.
That's a fair argument honestly I can see it going both ways. Offsite backups are more reliable but it's at the cost of downtime. It's really up to the provider on what they want to prioritize.
Yeah... Once again not sure where you are getting this from. Here is a list of ~500 Processors that have a dual setup: https://www.cpubenchmark.net/multi_cpu.html
I still have some IBM X3500 M4's in my garage I can take a video of and show to you running with both 1 and 2 processors.
By definition they may differ but in the real world they can be related. Lets take a RAID1 Array for example.
RAID 1 consists of an exact copy (or mirror) of a set of data, so basically one drive is a "backup" of another drive, in the event of a single drive failure, it is "redundant" as the "backup" drive remains operational for the system.
I might have gotten a little off topic with "backups" but I'm still speaking to a web-hosting deployment. A backup alone will result in downtime, redundancy + backups wont, to an extent.
- Status
