On Back-ups and our policy
Posted: Wed Feb 17, 2016 2:03 am
Hello!
There were some tickets lately about the back-ups we take.
I think this is a misunderstanding, we do not take any back-ups unless otherwise noted, only on our premium products, even then, they are for the bare metal restore, NOT individual VPSes, so, they cover the case of RAID card or multiple disk failures, for example, but not if your OS update went wrong, hackers deleted your files due to poor security or 0day exploits.
We offer a self-managed service, which means the administration of your VPS, including taking back-ups, is your responsibility, you have full root access (of course, this does not apply to our shared hosting plans, we only discuss VPSes here). We do help, however, with tutorials and FREE FTP back-up space where you can store your important data, all our customers can order such plans for free (10 GB) or paid for larger plans at a very low cost.
A good backup policy is not method-specific, you can use various methods depending on your type of data and location/knowledge, but on schedule, number of copies and physical location where they are stored.
You must follow the rule of three: three back-ups, at three different moments, in three different locations, if your data is very important. We assume it is not a life and death situation otherwise you would not keep it online off-premises, so we offer 2 locations, for now, for the free service, but there are more locations for paid service, practically all our locations offer storage products, except Pune where BW is very expensive and would not be feasible to run a backup service.
In case of hacking, accidental deletion, botched upgrades and installs, it is obviously not our fault, but even hardware failure is not our fault, everything made by humans can break, and sooner or later, it will.
We have redundant storage, all services run on RAID of a kind or another, specialized storage units (SANs) which cost 1 million Eur a piece, FC infrastructure and whatnot, but, while there is a lower chance of dataloss due to hardware failure, this can happen, and, since 2011 when we opened these services for the public, we had 3 separate incidents, on one we managed to recover the data in some 8 hours, the others lead to permanent dataloss. If it is good or bad, you decide, but i think this is too good as people learned to rely on our prowess instead of taking care of their important data.
Nobody is God or Superman, there will be more such failures, we try to keep them at a minimum, but at the end of the day, we can offer various degrees of safety at different levels of cost.
We never lost data on our premium products as we take extra steps such as taking back-ups for them (albeit, keep in mind it can still happen, even if we have backups and a disaster recover policy in place), however, self-managed plan means you have the full responsibility for your back-ups, even when we take one too, that is an extra for premium products to get back quickly up and running, NOT a replacement of your own back-ups, after we restored our backups, you should restore yours as they will likely be taken at the right time and be more recent in most cases.
Bottom of line...
IF your data is very important, THEN:
1. Use redundant locations containing exact replicas of the data, so, if one service fails, the others take over with very limited, if any, downtime;
2. Take at least 3 backups, at three different moments, in 3 different locations;
3. VERIFY, if not daily, at least weekly, the status of your automated backups, by attempting a restore someplace and see if the data is restoring and working correctly.
Failure to do so can lead to a lot of anger, lost work, money, frustration and angry tickets in our billing panels.
Please try to avoid it, if you value your work and data, take your back-ups now!
There were some tickets lately about the back-ups we take.
I think this is a misunderstanding, we do not take any back-ups unless otherwise noted, only on our premium products, even then, they are for the bare metal restore, NOT individual VPSes, so, they cover the case of RAID card or multiple disk failures, for example, but not if your OS update went wrong, hackers deleted your files due to poor security or 0day exploits.
We offer a self-managed service, which means the administration of your VPS, including taking back-ups, is your responsibility, you have full root access (of course, this does not apply to our shared hosting plans, we only discuss VPSes here). We do help, however, with tutorials and FREE FTP back-up space where you can store your important data, all our customers can order such plans for free (10 GB) or paid for larger plans at a very low cost.
A good backup policy is not method-specific, you can use various methods depending on your type of data and location/knowledge, but on schedule, number of copies and physical location where they are stored.
You must follow the rule of three: three back-ups, at three different moments, in three different locations, if your data is very important. We assume it is not a life and death situation otherwise you would not keep it online off-premises, so we offer 2 locations, for now, for the free service, but there are more locations for paid service, practically all our locations offer storage products, except Pune where BW is very expensive and would not be feasible to run a backup service.
In case of hacking, accidental deletion, botched upgrades and installs, it is obviously not our fault, but even hardware failure is not our fault, everything made by humans can break, and sooner or later, it will.
We have redundant storage, all services run on RAID of a kind or another, specialized storage units (SANs) which cost 1 million Eur a piece, FC infrastructure and whatnot, but, while there is a lower chance of dataloss due to hardware failure, this can happen, and, since 2011 when we opened these services for the public, we had 3 separate incidents, on one we managed to recover the data in some 8 hours, the others lead to permanent dataloss. If it is good or bad, you decide, but i think this is too good as people learned to rely on our prowess instead of taking care of their important data.
Nobody is God or Superman, there will be more such failures, we try to keep them at a minimum, but at the end of the day, we can offer various degrees of safety at different levels of cost.
We never lost data on our premium products as we take extra steps such as taking back-ups for them (albeit, keep in mind it can still happen, even if we have backups and a disaster recover policy in place), however, self-managed plan means you have the full responsibility for your back-ups, even when we take one too, that is an extra for premium products to get back quickly up and running, NOT a replacement of your own back-ups, after we restored our backups, you should restore yours as they will likely be taken at the right time and be more recent in most cases.
Bottom of line...
IF your data is very important, THEN:
1. Use redundant locations containing exact replicas of the data, so, if one service fails, the others take over with very limited, if any, downtime;
2. Take at least 3 backups, at three different moments, in 3 different locations;
3. VERIFY, if not daily, at least weekly, the status of your automated backups, by attempting a restore someplace and see if the data is restoring and working correctly.
Failure to do so can lead to a lot of anger, lost work, money, frustration and angry tickets in our billing panels.
Please try to avoid it, if you value your work and data, take your back-ups now!