[dkhenry]

I almost feel like I am missing out on hating MongoDB with others, because for me it just works, and _nothing_ I have seen people complain about is really a problem for me. Even in this article your upset because if you have multiple threads using a single connection the getLastError message won't always report the query form the current thread ? That's like saying your upset because the linux kernel does not let other threads know your in the middle of a mutation of a memory address so it may get corrupted. Its your fault deal with it.

Then your upset because in the past it didn't preform like you wanted it to so somehow that should be factored into the _current_ reliability ? My takeaway is you didn't care enough about your data in the past to understand your persistence layer why do I feel you are going to do the sane thing in the present and understand your persistence layer now , it appears if it doesn't work like you want it to by default then its broken and 10gen should really be making a datastore just for you. I am more annoyed that I need to now specify that I _don't_ want to wait for confirmation of write, but I at least took the time to understand how the DB works and what I need to do with it.

[banachtarski]

What is your usage of it? Saying it works for you might mean nothing at all if you do 5 ops/second on average and have < 10 gb of data or something.

Non-gamers are perfectly happy with a shitty graphics card too.

[mrkurt]

Does it matter? We do a pretty continuous 400k ops/s worth of Mongo, but I don't think that's necessarily going to fix anyone's issues with it.

[moe]

400k ops/s?

How many machines is that?

[dkhenry]

I can do it one one.