Per http://www.spry.com/webmin-vps/, Spry will do "Full nightly backups are included with all VPS plans. Not only do we backup your website, we backup your entire operating system."
In the recent past when I had a Spry tech rebuild my VPS with another OS(which he was very helpful in doing without a problem - Way to go Spry!) I asked the tech to dump the latest backup of the old system to /old on my newly rebuilt VPS. The tech indicated that all he could grab were the last couple days of incremental backups. When I asked him to also include the latest full backup he said that SPRY only keeps 1 week of backups. The problem with that, in my case, is that none of them were full backups so all I had of my data was a week's worth of new stuff, but none of the GBs of old data.
What is the real SPRY backup policy? Full nightly backups of data+OS as advertised on teh site, or the week of incrementals? Either is perfectly fine with me (my belief is : if one really needs a backup of one's data, one needs to make MANY copies of it oneself), I just need to know which one it is so I can plan out my own backup / retrieval strategy.
Last edited by portalgod; 11-14-2007 at 10:46 AM.
According to our backup system administrators documentation, our backup system does full backups once a week. Every other evening, an incremental backup of all changes since the last backup operation is performed.
Originally Posted by portalgod
Due to the vast amount of data that is all Spry VPS's (OS and data) currently totalling over a couple terabytes, our backups sytem takes several hours to restore files. It is also not immune to temporary failures, preventing our technicians from being able to access the last full backup.
Our backup system is designed with catastrophic hardware failure in mind, meaning its main purpose is to ensure we are able to recover hosted VPS's in the event of a hardware node failure. In this case, only the most recent incremental backup is required (since most of the data will still be accessible on the hardware node) and all files for a hardware node would be extracted in one restore operation.
If you need to retain more than a couple days worth of backups, we highly recommend configuring your own backup system/strategy without regard to our backup system. There are several advantages to doing so:
* Our restores take a long time due to the huge amount of data that must be sorted through to extract your files. (Several hours in most cases.)
* You can control how your backups are run and can make incremental backups more often than we are able to. You could, for example, run an rsync two or three times a day rather than once nightly as we do.
* You can lock important files before a backup to make sure they are backed up into a consistent state, for example your MySQL files could be locked before a run and then unlocked afterwards. We don't know which files our customers would need treated this way, so the dump just captures all the files as they run, and they can change while being dumped. (Some MySQL database storage engines are highly susceptible to data loss if not locked prior to backup via file system dump.)
* You can keep your backup files around for as long as you want.