[baltimorepm] Seeking advice for using Perl with NoSQL

[Stephen Belcher]

> I initially considered using MongoDB, but I've read it has data integrity
issues under certain conditions.

I'm not very experienced with MongoDB, and particularly not to the point
where it crashes and/or burns, but my impression from anecdotes of same has
been that the corruption tends to happen under very high-volume conditions,
which may or may not be your use case. But, though I've heard of this a few
times with MongoDB, I think this is a problem that could occur in many/most
NoSQL databases, as they often trade off some data durability for a speed
boost since disk I/O is frequently the worst application bottleneck, and
large-scale applications needing that kind of extra speed is frequently
their target market.

> All I know right now is that the data I need to handle comes in all
different forms and values

Another database you may be interested in checking out is Riak (
http://basho.com/riak/). Depending on whether you meant "different forms
and values" to mean different -- potentially binary -- file formats/etc.,
it may suit your situation better than MongoDB, which requires you to
encode your data as JSON. Additionally, Riak has an HTTP interface for
doing create/read/update/delete, and since it takes whatever data/files you
throw at it, the barrier to entry is low from an API standpoint. However, a
coworker trying to set it up did complain of a size limitation on
individual objects of around 50 megs which maybe can't be configured
around; but again, this may not be applicable to your use case, or be based
on incorrect/outdated information on my part.

> Are my fears re: Mongo overblown?  Is anyone happy with using something
else?

If, on the other hand, you're already emotionally invested in MongoDB, and
just want someone to better assuage your concerns/convince you it's the
best fit, there's been rumblings on the DC list about Mike Friedman from
10gen coming to give a talk on using Perl with MongoDB. It should be very
informative, as he's the author of the driver on CPAN. And, according to
http://dc.pm.org/Home#topicqueue, it's currently scheduled for September
3rd. So, depending on your patience towards both waiting and traveling to
DC, that's worth a look.

Hope this helps,
--Stephen
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.pm.org/pipermail/baltimore-pm/attachments/20130801/569fc7b0/attachment.html>