This stuff definitely looks pretty raw. Mnesia seems to be quite mature in comparison, but isn't the greatest option for a system not fully written in Erlang.
Has anyone tried this Project Voldemort referenced in the post? Skimming the description on the project page, I get the impression that it's fairly well thought-out.
"Follows the bigtable model, so it's more complicated than it needs to be. (300+kloc vs 50 for Cassandra; many more components). This means it's that much harder for me to troubleshoot. HBase is more bug-free than Cassandra but not so bug-free that troubleshooting would not be required."
I only had problems (data losses) when I was using HBase in production 3 months ago, but I'm certain that HBase will be mcuh more stable in a few months. And there comes that Cassandra developer and says that Cassandra even has more bugs. Lol. Now I'm pretty sure that I will never touch Cassandra as well.
Anyone have a good dataset I can use to populate these things and give them a workout? It'd be cool to get millions / billions of records of something real that I could test with. I could always generate something but I like real data better for such business.
Has anyone tried this Project Voldemort referenced in the post? Skimming the description on the project page, I get the impression that it's fairly well thought-out.