post about MongoDB that made it all over the Net this weekend thanks to being voted up on Hacker News.Mark Twain said, "a lie is halfway around the world before the truth can get its boots on." The speed of putting on boots (if you wear them) hasn't changed much since Twain's day, but lies and misinformation are getting a serious boost out of the Internet. Consider the
In short, the anonymous Pastebin post slams MongoDB saying "I now feel a kind of social responsibility to deter people from banking their business on MongoDB" and then lists reasons why MongoDB is a bad choice. Sounds like typical HN front-page material, right?
The problem is, the veracity of the original piece is in question. True, false, or a mix of both – it made it to the front page of Hacker News very quickly. Who knows how many folks have read it since then? Certainly a lot more folks than have read the comments thread or the follow-up from the poster claiming "it is in fact a hoax that has gone too far, you got trolled."
To be fair, a lot of the people on the HN thread don't seem to have taken the post at face value. There's also been effort to spread the response from 10gen alongside the original post.
A Grain of Truth?
To 10gen's credit, Eliot Horowitz (CTO) was in the thread and responding to the post right away on a point-by-point basis. Even more, Horowitz respected the forum and wasn't spinning the usual corporate response. Even though Horowitz had clues the poster was trolling (the original post has something about having a "crazy platinum support contract," which Horowitz debunks), he did acknowledge some of the actual problems that exist or have existed with MongoDB.
I also spoke with Horowitz and 10gen president Max Schireson yesterday about the post and problems with MongoDB. Again, rather than deflecting the entire thing, Schireson and Horowitz were fairly candid about MongoDB's shortcomings. Of the nine sections, Schireson says that "some are definitely valid, some we haven't heard or seen."
For instance, Schireson says that the complaint about the global lock is a "well-known deficiency," but "something we're working on to improve."
The biggie for any database, of course, is data loss. Is that an issue? Not anymore, says Schireson, but there were issues prior to 1.8. "If you only used a single server, then there were reliability issues where you might lose data. We tried to clearly document that, and make clear that the recommended approach was to use multiple servers to achieve reliability." Since 1.8, says Schireson, "we do not believe there have been substantiated reports of data loss when the system is used in the recommended way."
Perhaps surprisingly, Schireson didn't seem too upset with the post. "I tend to rely on users to be pretty smart and sophisticated." Schireson admits that it "created some FUD for us." But, looking at the response and comments voted to the top, Schireson says that most users seemed to be examining the issue pretty carefully.
So should users avoid MongoDB? I'd say they should take the same approach as the HN skeptics: evaluate it and find out for yourself. It does sound like MongoDB, being a relatively new project, has some sharp edges. But given some of the success stories (such as Foursquare), it's clear that MongoDB can also be used successfully in production for large data sets.
The System Works?
Sites like Hacker News are a blessing and a curse for dissemination of news. The good news is that they can provide a platform for information and viewpoints that wouldn't otherwise be heard. (Though HN these days is less and less non-mainstream.) The bad is there's little editorial quality control.
A single-sourced, anonymous, piece like this one wouldn't (or at least shouldn't) be making the front page of widely read publications.
But the reaction on HN is heartening. Though it's disappointing it made it to the front page at all, it also seems that the bulk of the audience at HN took it with the grain of salt it deserved.