I have been looking at ways to keep a lid on the potential size of a mySQL DB in the webapp that I plan to get into beta soon. The DB has the potential to become very big. Having examined what I am doing I found that using the file system to store certain data (mostly images) would have a huge impact on DB size - ensure that I would probably never have to deal with a DB that blows the TB ceiling though it might still get into 100s of GBs. I ended up discovering TokuDB in my quest for the answer to the question - "would it be worth compressing my VARCAHAR data prior to putting it into the DB?".
The benchmarks I have seen for TokuDB are very impressive. However, I have also run into a few comments that are less than reassuring
- It is slow on ON DUPLICATE UPDATES. This has been countered by another writer who claims that a simple switch to REPLACE INTO fixes the issue.
- It is good enough for a backup DB but not for production purposes.
Those comments are more than a year old so I was wondering... what is the current view? I would like to hear from anyone who has experience with using TokuDB. The tables on which I may use it have a high proportion of write accesses and typically have one largish VARCHAR column.
