Ian and I are on the trail of a possible misconfiguration of the Redis installation on cpsc. While we diagnose that, let me give you this piece of advice (which is sensible anyway, regardless of this problem): if you’re going to be entering data in bulk (say, by typing lots of Pokémon stats, or lots of fishing spots, or lots of crochet patterns) do so in a Python file, rather than entering them interactively. In other words, write a .py file with lots of insertions, like this:

r = redis.Redis(db=43, password="YouKnowWhat",
    decode_responses=True)
r.hset("hitPoints","charmander","45")
r.hset("hitPoints","pikachu","35")
r.hset("hitPoints","ditto","55")
r.hset("hitPoints","squirtle","25")
    . . . and on and on and on . . .

Then, if something goes south on the database, and all the Redis data is lost, you just say “no sweat; I’ll just re-run my Python program and populate my Redis instance with its initial contents again.” If, by contrast, you had actually typed all those Pokémon stats in at the redis CLI, you’d be grumpy if they disappeared and you had to type them all over again.

Stay tuned for more info on the mysterious data loss bug.