2 ms·
If using `db-s3-object`, does KeyDB uses that as a seed on startup? Eg: Running KeyDB on AWS, the EC2 instance fails and is replaced by another one running the
by xfalcox 8y ago
If using `db-s3-object`, does KeyDB uses that as a seed on startup?
Eg: Running KeyDB on AWS, the EC2 instance fails and is replaced by another one running the same config. Will it read the last S3 dump on startup?
- jdsully 8y agoNot yet implemented. I’m hoping to get that fixed today. I mostly use slaves for recovery with the database dumps as a last resort. So I overlooked that.
- xfalcox 8y agoAmazing! Do you have any special config to sync to replica before accepting incoming connections in a straightforward manner?
- jdsully 8y agoYea we have external management software we wrote. We’ve been looking into releasing this as well, but it’s not cleaned up yet.
- xfalcox 8y agoLooking forward to that. In the meanwhile I'll wait and test the S3 option when it's live.
- jdsully 8y agoI just implemented S3 loading in this change: https://github.com/JohnSully/KeyDB/commit/a45f212693956a6fb1aacf465d88e940bbbfd56f https://github.com/JohnSully/KeyDB/commit/a45f212693956a6fb1... Its currently in the unstable branch, once it gets shaken out a bit I'll release 0.9.3 with it included.
- lessclue 8y agoThis is a nice project. Hardcoding "aws" in the core seems like an odd choice though. Wouldn't it be better to make it agnostic and provide some sort of a trigger where a an external script or util to handle backups? That is, why S3 and why not any other service?