Technical writing

Cassandra
randomness
& deep dives

Field notes from running Apache Cassandra in production — compaction strategies, cluster migrations, query debugging, and the bits nobody documents.

7 Articles
5+ Years exp.
93 Clusters migrated
All posts RSS feed →

Cassandra TWCS must have TTLs

Time Window Compaction Strategy is powerful for time-series data — but only if you follow the rules. Here's what happens when you don't attach a TTL.

Moving Cassandra Clusters without Downtime – Part 1

Setting up a new datacenter and migrating Rackspace clusters to Google Cloud, zero downtime. Part 1 of 3.

Moving Cassandra Clusters without Downtime – Part 2

Decommissioning the original Rackspace datacenter after successfully migrating data to Google Cloud. Part 2 of 3.

Moving Cassandra Clusters without Downtime – Part 3

Things that can go wrong during a zero-downtime Cassandra migration — and how to mitigate and recover. Part 3 of 3.

Finding Rogue Cassandra Queries

Using ngrep and Wireshark to catch queries running at the wrong consistency level — proving it to the development team.

Cassandra Counting without using Counters

Why count(*) on large Cassandra partitions is hurting your cluster, and a smarter pattern using timeuuid and lightweight transactions.

Who is Connecting to a Cassandra Cluster?

Tracing unexpected cross-DC connections using ngrep on port 9042 — finding which clients are ignoring local consistency.