ReadMeFeed / Community knowledge
Real questions. Useful conversations. Find the people who know your stack.
Frameworks & web · #hadoop
The default HDFS replication factor of three is a guideline, not a mandate. Balancing data durability against storage cost and write latency requires matching the replication factor to your specific failure domain tolerance.
Reduce HDFS storage costs by applying Reed-Solomon erasure coding to cold archival directories while keeping hot workloads on three-way replication.
A practical guide to identifying and resolving under-replicated blocks and block report mismatches in Hadoop 3.x HDFS clusters.
Learn how Hadoop MapReduce speculative execution works, when to adjust its wait time, and what trade‑offs to watch for on a shared cluster.
When HDFS stays stuck in safe mode, writes fail and jobs stall. This guide walks you through the root causes, diagnostic checks, and safe‑mode exit strategies to get your cluster back online.
Triple replication costs 200 percent overhead for cold HDFS data. Erasure coding in Hadoop 3.x drops that to roughly 50 percent, but adds read and reconstruction cost. Here is how to decide, configure, and verify.
Cut HDFS storage costs with erasure coding: apply a 4+2 policy to a directory, write a 500 MB file, verify 1.5× overhead, and understand trade‑offs like recovery latency and write‑once restrictions.
Stop wasting 200% extra storage on cold data. Learn how to implement HDFS Erasure Coding to reduce overhead to 50% while maintaining fault tolerance.
A diagnostic guide for resolving HDFS Safe Mode persistence. Learn how to identify block health issues, verify DataNode connectivity, and safely exit read-only mode.