Hi, I'm going to count activity.
Is it good Idea to use HBase to count?
Count it cause hotspots?
What are the optimisations for HBase to increase counters throughput?

I've found old letter here:
http://search-hadoop.com/m/xfABgvwuvQ1/counter+increment&subj=Re+on+the+impact+of+incremental+counters

Does it make sense to increase WAL flush to avoid it be too frequent? I
understand that I can loose the data if WAL is not flushed and RS failed.

Reply via email to