Configuring the background verification rate

You can change the rate at which background verification checks replicated object data on a Storage Node if you have concerns about data integrity.

Before you begin

About this task

You can change the Verification Rate for background verification on a Storage Node:
  • Adaptive: Default setting. The task is designed to verify at a maximum of 4 MB/s or 10 objects/s (whichever is exceeded first).
  • High: Storage verification proceeds quickly, at a rate that can slow ordinary system activities.

Use the High verification rate only when you suspect that a hardware or software fault might have corrupted object data. After the High priority background verification completes, the Verification Rate automatically resets to Adaptive.

Steps

  1. Select Support > Grid Topology.
  2. Select Storage Node > LDR > Verification.
  3. Click Configuration > Main .
  4. Go to LDR > Verification > Configuration > Main.
  5. Under Background Verification, select Verification Rate > High or Verification Rate > Adaptive.

    setting Verification Rate
    Note: Setting the Verification Rate to High triggers a Notice level alarm for VPRI (Verification Rate).
  6. Click Apply Changes.
  7. Monitor the results of background verification.
    1. Go to LDR > Verification > Overview > Main and monitor the attribute Corrupt Objects Detected (OCOR).
      If background verification finds corrupt replicated object data, the attribute Corrupt Objects Detected is incremented. The LDR service recovers by quarantining the corrupt object data and sending a message to the DDS service to create a new copy of the object data. The new copy can be made anywhere in the StorageGRID Webscale system that satisfies the active ILM policy.
    2. Go to LDR > Erasure Coding > Overview > Main and monitor the attribute Corrupt Fragments Detected (ECCD).
      If background verification finds corrupt fragments of erasure coded object data, the attribute Corrupt Fragments Detected is incremented. The LDR service recovers by rebuilding the corrupt fragment in place on the same Storage Node.
  8. If corrupt replicated object data is found, contact technical support to clear the quarantined copies from the StorageGRID Webscale system and determine the root cause of the corruption.