On Mon, Jul 14, 2014 at 8:00 AM, John Mazzitelli <mazz(a)redhat.com> wrote:
> First, we are dealing with append-only data, where each column is
only ever
> written once and never updated.
Will this always be the case? We do plan on using C* for future uses - I'm wondering
what happens if we do have data that will not be "append-only" in the future?
Should we leave code in here to prepare for that? Or will we always ensure any data we put
into C* fits this model of append-only (written once, never updated)?
I do wonder about this supposed 'future use' as I was hoping RHQ would
push for more data in C*. :-)
As for traits and call times, for the patch I put together anyway, I
use TTL for deleting old data. Exception: When deleting a resource,
deletion is done by schedule ID and it is possible for this delete to
be lost between nodes. I'm not sure this would warrant a weekly job.