Erik Krogen
Erik Krogen
Though Dynamometer supports executing a few different versions of Hadoop, it will currently only run on a 2.7 cluster. This tracks adding support for the rest of the 2.x line.
This fixes the issue described in PR #77 regarding improper parsing of the FSImage XML when the CacheManagerSection is present.
The block generation job has custom output logic to allow each reducer to output to multiple block files. When speculative execution is enabled, this can result in two copies of...
The Hadoop tarball to use during testing is specified and downloaded from an Apache mirror. Generally only the latest version of each branch is available, but right now, versions must...
Though the project has some tests, it is essentially just one monolithic integration test. We need more fine-grained unit tests to be able to more easily diagnose issues as they...
### What changes were proposed in this pull request? This modifies `GenerateUnsafeProjection` to wrap projections of non-null fields in try-catch blocks which swallow any `NullPointerException` that is thrown, and instead...
Primarily closely read the paper's description of the algorithm and made changes to match the spec to the text description. One of the biggest changes was that nodes get marked...
When using a data lake connector (Hive/Iceberg/etc.) to write data to HDFS using Trino, we may see a [QuotaExceededException](https://hadoop.apache.org/docs/current3/api/org/apache/hadoop/hdfs/protocol/QuotaExceededException.html) (e.g. namespace quota or disk space quota exceeded). This is a...