Resilience of Archive Storage


#1

I’m currently reviewing my backup/recovery strategy for my VM. As part of this (only part I hasten to add) I have archive level storage on my VM and backup data to there. In the event of a problem my first port of call would be to recover from here.

I understand that the archive storage is separate from the main storage but I seek some confirmation.

If I make a mess of my VM and re-image will my archive storage still be available?

If Bytemark have a major problem and my main disk is totally lost can I recreate a VM with access to the archive storage?

Regards

Tony Middleton


#2

Hi Tony,

Yes, its sort of separate, apart from Archive storage being physical hard disks versus the SSDs being (unsurprisingly) SSDs.

The way the Cloud Server platform is set up means that the ‘tails’ which contain the disks allow us to transparently move virtual ‘disks’ between physical servers - these have multiple volumes set up which are effectively groups of either Archive or SSD storage, and while there are a number of pools over a number of physical servers, there’s a chance both and Archive disk and SSD may be on the same physical server at some point.

Re-imaging a server basically creates a new root disk of the relevant size and deploys the disk image to it - it doesn’t touch the configuration of the server, so any archive disks should be untouched although they will need to be remounted of course.

It’s important to note though that deleting the server will delete all the disks attached to it, and it’s not possible (at least at the moment) to attach an archive disk to another Cloud Server.

Finally, it’s always good practice to have off-site backups, so if you’re really looking for resiliency in the event of something going wrong, you’d want a separate copy of any backups which aren’t generally accessible by the server (either in a different location, and/or hosted with another company) as if the any disk is mounted, then an rm -r / has a good chance of still removing local backups!


#3

In case it is helpful…

My server RSYNCs the automatic backup to my home NAS drive every night. This means I have a back up of all my SQL databases.

I also rsync (via a cron job) my /srv folder to my home NAS driver every night.
rsync -azP --exclude 'logs' --exclude 'Maildir' --exclude '.webalizer*' --exclude 'cache' --exclude '.well-known' --delete /srv/ ****@*****:/volume1/homes/*****
This means I get an email with a list of all the files that have changed. Useful to make sure nothing odd is going on (files changed by a hack or anything)

My NAS drive then takes a daily snap shot of the target folder for backups, resulting in me having a daily backup history of all my files and databases.