Skip to content

Allow restart operations when aborted or failed #11

Description

@juarezr

Operations in backup or restore can be interrupted by:

  1. Errors in source docs:
    • Documents without required fields
    • Corrupted documents
    • Generally when user has changed schema after imported docs
  2. Server/Cluster problems:
    • Server overload
    • Huge transaction log when restoring (avoided by using --commit hard)
    • Replication throubles
    • Network problems between cluster instances
    • Out of memory in JVM / Solr
  3. Interruptions:
    • Power failure, computer restart, process killed
    • Lost ssh connection
    • can avoided using the screen command before running solrcopy

Todo:

  1. Create a mecanism for recording checkpoints for backup/restore:
    • Backup can require using the --order for checkpointing
    • Restore already has --order using zip filename that could be extended for checkpointing
  2. Create a new switch like --continue that could:
    • detect when the command failed before
    • restart from the point interrupted
    • can skip in case of invalid/corrupted docs

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    enhancementNew feature or request

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions