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.
To be Clear! Scaleway is unsecure! Your data could fail anytime...
Hey,
I wanted to start a serious project with scaleway. I opened a support ticket about data security.
To BE CLEAR...
They use the dedibox XC without any raid configuration. So one disk/server. Also the LSSD node where you got the volumes not use any kind of raid.
They said your data could fail anytime as they not use any redudancy. (i found that in the faq)
Conclusion:
Not really matter you use the online.net dedibox XC or scaleway , as you got the same data security. This just a hour-based servers nothing more or nothing special.
Comments
I don't think they make any attempt to hide this. I also don't understand why people buy servers without RAID, but the fact is they do. I guess the risk is worth the cost saving in some people's eyes...
What do you expect from low-end provider?
Backup, backup and backup.
Backup not a problem. But configure everything every time is another story.
If they dont use raid at all how they setup their network drive? Just put bunch of drive and slice them per 50 GB?
I imagine they create LVs on demand and export them over NBD.
@ZweiTiger - you know you can back up configuration files too, right?
They offer snapshots for local backups/easy re-deployment. https://www.scaleway.com/docs/backup-your-data-with-snapshots/
You should always have offsite backups for critical data.
Yeah.. more work more time.
Regarding the support ticket yes. Thats all
You have to shut down the server to do that. So... not a good option
Not necessarily. You could use for example Idera CDP ?
4/ We do not provide hardware RAID. One way to protect against hardware failure is to use software RAID or backup your files on Scaleway SIS our object storage solution.
Refer to this: https://community.scaleway.com/t/i-have-questions-about-scaleway/891/2
So get two nodes and setup redundancy between them. You could either rsync everything over and over again or be creative with drbd or try to make something like RAID1 over iscsi - there are multiple options. Having 2.5Gbps network helps.
For some uses a box with a single disk / no RAID is useful and good enough.
This is the main reason I still stay with leaseweb and prometeus IWStack. Almost redundant setup from network to storage. But they may not fix into LET price for higher end.
Why not use ansible?
At least they are clear and honest about it. How sure are you all that your current provider is really using raid.
Regardless of what you think they use or claim the position you should always be in is having a restore point available that is no more than 15 minutes older than a nodes failure time. And you should have the expectation that reconfiguration at a new provider will be required and plan for that accordingly.
Not know it , but seesm their price is high.
I used iuera for 1 year. Not really like it.
Stating tolkien "Stepping out of your own door" is unsecure and rather dangerous. If you run a production Server there make proper backups to external servers and you will be fine^^
Ansible is free, Ansible Tower is paid.
Snapshots seem to be a secure option, at least that's how it sounds in their faq. They are not kept on the same LSSD.
That's for the tower, you don't need that. Check out: http://docs.ansible.com/ansible/intro_installation.html
I have a friend who is fed-up with RAID for his small business. They have had so many issues and lost data. I can see his point of view, although he is not 100% right about it, either.
I am looking into this at the moment. Why do you use Ansible vs puppet, salt, chef, cfengine and all other that exists?
This is know, drop them a ticket to let you know if the second volume is on another Storage thingy. If yes, they will fix it.
Then you actually have better redudancy as Raid 1 local.
Same thing als DO or Vultr, you could be on the same Node when you create 2 VM's.
It was one of the first and easiest I found
Well, I would say "You get What you Pay" and Thats the universal truth fr the Industry.
But, Then RAID or NO RAID, You must not forget the seond Universtal Law : YOU MUST BACKUP 24x7 ! ( BurstNet can happen any time )
So,
Firstly, I feel Any plan from them coupled with a decent backup plan , hosted at Time4 is not really bad for most needs. ( Remember we are discussing this at LET , Price matters above all ).
Secondly, If you are serious about your project , Start with ScaleWay and when you grow....then why not get a decent VPS with RamNode with most reliable and affordable plans.
That's why they do have snapshots and, even better, creating own images from snapshot. They do keep the snapshots and images in different disks and config than the LSSD, so, if your "local" disk crashes, you can recall at anytime the provided own image from a snapshot and save time and effort to setup all things again.
Preview/Beta product.
Found your problem!
RAID is not a backup, security function, or anything of that nature. If you rely on RAID for such reasons you are doing IT horribly, horribly wrong.
Your data could fail anytime with RAID, fyi.
Pretty much the only situation where you can treat data as "safely stored" is when it has completed the round trip to replicate between Datacenter A, B, and an encrypted backup has been stored in a 3rd datacenter with a separate provider.
Can't you have multiple 'disks' then run software RAID on them within the instance?
Yes, but your "disks" might be all on the same physical disk