Back then they seemed to have a rather sensible migration strategy (ie, basically running the new Cassandra back-end in parallel with the MySQL backend).
It seems to me that it was the v4 upgrade that broke, not Cassandra alone. It's possible their frustrations with Cassandra were more long term, and the fact the v4 upgrade didn't go well was the last straw.
For example, Digg has done a lot of work on Cassandra internals and tools. If you are using a new, open source product you kind of expect that, but it's possible the expense of that didn't seem like good value once v4 started to get into trouble,
Back then they seemed to have a rather sensible migration strategy (ie, basically running the new Cassandra back-end in parallel with the MySQL backend).
It seems to me that it was the v4 upgrade that broke, not Cassandra alone. It's possible their frustrations with Cassandra were more long term, and the fact the v4 upgrade didn't go well was the last straw.
For example, Digg has done a lot of work on Cassandra internals and tools. If you are using a new, open source product you kind of expect that, but it's possible the expense of that didn't seem like good value once v4 started to get into trouble,