7.2.20 scrub

On-demand Exascale file scrubbing for logical consistency.

Syntax

scrub [ -a ] [ -i interval] [ -n requests] [ -r count]
      [ -f offset] [ -t offset] [ -s ] [ -v[v[v]] ] 
      file-name
      [{ -w | --wallet } wallet-location ] 
      [{ -T | --trace } trace-level ]

Command Options

The options for the scrub command are:

  • file-name: Specifies the files being scrubbed.

    You can use a wildcard (%) in the file-name to specify multiple files.

  • -a: Specifies that the scrub operation performs HARD checks on all regions, regardless of the file type or mirror status.

  • -i: Specifies the interval (in seconds) to wait before re-scrubbing a region.

  • -n: Specifies the queue depth for the scrubbing operation. This option controls the number of concurrent read requests performed by the scrubbing operation. Larger values enable faster scrubbing but consume more resources, which may affect other workloads.

  • -r: Specifies the number of re-scrubbing attempts permitted while awaiting concurrent writes on the region. After the specified number of re-scrubbing attempts is exhausted, the corresponding region is reported in the command output and the scrubbing operation continues.

  • -f: Enables scrubbing on part of a file by specifying the file offset that scrubbing starts from. The offset value is specified in megabytes from the start of the file.

  • -t: Enables scrubbing on part of a file by specifying the file offset that scrubbing operates to. The offset value is specified in megabytes from the start of the file.

  • -s: Specifies that when considering multiple sets of data as the source for a repair, the scrub operation should use the highest system change number (SCN) as the repair source.

  • -v[v[v]: Increases the verbosity of the command output:

    • -v: Generates detailed output.
    • -vv: Generates more detailed output.
    • -vvv: Generates the most detailed output.
  • -w, --wallet: Optionally specifies the path to the Exascale wallet directory.

  • -T, --trace: Optionally enables tracing, with the trace level (trace-level) set to 1 (minimum tracing), 2 (medium tracing), or 3 (maximum tracing). If the trace level is not specified, then minimum tracing is enabled by default.

    The trace file is written to the first accessible location in the following list:

    1. If the $ADR_BASE environment variable is set:

      $ADR_BASE/diag/EXC/xsh_<user-name>/<host-name>/trace/xsh_<date>.trc
    2. /var/log/oracle/diag/EXC/xsh_<user-name>/<host-name>/trace/xsh_<date>.trc
    3. /tmp/diag/EXC/xsh_<user-name>/<host-name>/trace/xsh_<date>.trc

Usage Notes

Logical file scrubbing is a valuable tool for addressing certain issues that may affect the integrity or consistency of Exascale files. However, regular or periodic scrubbing is not required or recommended.

Examples

Example 7-35 Scrub an Exascale file

The following example performs logical scrubbing on @my-data/my-file.

$ xsh scrub @my-data/my-file

Related Topics