Is Glacier for Backups part 2

In my previous post, I shared my opinion on whether Amazon Glacier should be used for backups. I concluded at the time that while it can be used for that purpose, there were significant downsides to doing so. The main one being that old data can’t be easily pruned nor overwritten.

There is an alternative way of getting your data up onto Glacier, avoiding the Glacier API and retaining the ability to prune overwrite and remove your files easily. That way is to upload your data to S3 and use it’s Lifecycle management to transition your data to Glacier. By backing up your data this way, you gain all the benefits of storing your data in S3, along with the cost benefits of Glacier.

Arq has started doing this in version 4 of it’s app and I have to say it’s changed my mind about backing up to Glacier. You can now overwrite and remove data in Glacier as easily as it was stored in S3, because it actually is, it’s just stored with the Glacier storage class. No more dealing with Vaults and Archives.

So now for the big question, would I now backup my data to Glacier? Yes. As long as your backup tool backs up to Glacier via S3, and not via the Glacier APIs.

Is Glacier for backups

Amazon Glacier is an amazing way to store data you no longer need, but want to keep just in case you might some day. Amazon promotes Glacier as an archival service, although some have been using it to backup their data. But should you? It's … [Continue reading]

Apache, Akamai, and True-Client-IP

You're running a site behind Akamai, but can't see the real client's IP address, only that of your load balancer and/or Akamai's IP addresses. Therefore you can't act on it, ban IP's....etc. Akamai does pass the true IP address of clients, but not … [Continue reading]