fcrepo-doctor

A tool for diagnosing and repairing OCFL-based Fedora repositories.

Requirements

Java 11

Usage

Usage: fcrepo-doctor [-dhvV] [--disable-auto-versioning] [-f=<problemsFile>]
                     [-o=<outputDir>] [-p=<parallelism>] -r=<ocflRoot>
                     [--s3-access-key=<s3AccessKey>] [--s3-bucket=<s3Bucket>]
                     [--s3-endpoint=<s3Endpoint>] [--s3-profile=<s3Profile>]
                     [--s3-region=<s3Region>] [--s3-secret-key=<s3SecretKey>]
                     -t=<tempDir>
A tool for diagnosing and repairing OCFL-based Fedora repositories.
  -d, --debug                Enable stack traces
      --disable-auto-versioning
                             Disables auto-versioning, which means that
                               applying fixes will NOT automatically result in
                               a new version of the affected resources.
  -f, --fix=<problemsFile>   Path to a file that contains a list of files to
                               fix. This file is generated by first running the
                               command without this option.
  -h, --help                 Show this help message and exit.
  -o, --output=<outputDir>   Path to a directory to write output files into.
                               Default: current directory
  -p, --parallelism=<parallelism>
                             Number of threads to use. Default: number of cores
                               minus one
  -r, --ocfl-root=<ocflRoot> Path to Fedora's OCFL storage root. When using S3,
                               this is the prefix within the bucket that the
                               storage root is located it, and it should be an
                               empty string if no prefix is used.
      --s3-access-key=<s3AccessKey>
                             S3 access key. If provided, a secret key must also
                               be specified
      --s3-bucket=<s3Bucket> S3 bucket the OCFL repository is in
      --s3-endpoint=<s3Endpoint>
                             URL to the S3 endpoint
      --s3-profile=<s3Profile>
                             S3 profile to use.
      --s3-region=<s3Region> S3 region
      --s3-secret-key=<s3SecretKey>
                             S3 secret key. If provided, an access key must
                               also be specified
  -t, --temp=<tempDir>       Path to a directory on the same filesystem as the
                               OCFL root to use for temporary files
  -v, --verbose              Enable more verbose logging
  -V, --version              Print version information and exit.

fcrepo-doctor is intended to be run in two phases. The first phase scans a repository and writes a fcrepo-doctor-problems.json file that contains all of the resources that were identified with problems. This first phase can be run while Fedora is operating normally.

In the second phase, the fcrepo-doctor-problems.json file is passed back to fcrepo-doctor, and it attempts to resolve all of the problems that identified in the first phase. For this phase, Fedore MUST be shutdown or at least not be actively written to.

Resources that are not fixed successfully are rolled back and logged to fcrepo-doctor-failed-problems.json. It is safe to stop the utility while it's running. It will gracefully terminate, and write fcrepo-doctor-incomplete-problems.json, which contains all of the problems that have yet to be fixed. This file can then be passed back into the utility to resume where you left off.

Example

# Phase 1: Analyze the repository
java -jar fcrepo-doctor.jar \
  --ocfl-root fcrepo/data/ocfl-root \
  --temp fcrepo/data/ocfl-temp
  
# Phase 2: Fix any problems found
# You may want to tee these logs to a file
java -jar fcrepo-doctor.jar \
  --fix fcrepo-doctor-problems.json \
  --ocfl-root fcrepo/data/ocfl-root \
  --temp fcrepo/data/ocfl-temp

Problems

Invalid Binary Description RDF Subjects