X hits on this document

PDF document

Can you really recover your data? - page 15 / 17

38 views

0 shares

0 downloads

0 comments

15 / 17

Ingres specifics

  • Database

    • Checkpoint Frequency does not affect RPO.

    • Backup frequently to reduce Journal recovery time.

    • Utilise high speed devices and/or parallel checkpoints.

  • Checkpoints

    • Must have more than 1 valid checkpoint and all its journals.

  • Dump Files

    • Backup with the checkpoint.

  • Journal Files

    • Backup at RPO-1 frequency.

  • Database Config File (in the dump directory)

    • Backup at RPO-1 frequency.

  • Transaction Log File

    • Regular consistency points.

www.Comp-Soln.co.uk

Copyright ©2005 Comprehensive Solutions International Ltd.

All Rights Reserved

Ingres specifics

  • RPO Factors

    • Archiver Interval must be less than RPO

      • cp_interval may not be good enough with large log file. Use cp_timer

      • r3. Possibly reduce archive_refresh and/or cp_interval_mb

      • Possibly force consistency points (dm1305) and Archive cycles (dm1314)

    • Journals must be copied when consistent, i.e., Archiver is inactive

      • Monitor iiacp.log

      • Check journal size before and after copy

        • Or check for Archive Cycles during the copy

      • Consider LOCAL (fast) and REMOTE (slow) archives

        • CCSI Tools: JournalArchive.ksh

      • Monitor journal archive process to ensure desired frequency.

    • Config file must be backed up (from Dump location) every time it is backed up by Ingres.

www.Comp-Soln.co.uk

Copyright ©2005 Comprehensive Solutions International Ltd.

All Rights Reserved

29

30

15

Document info
Document views38
Page views38
Page last viewedMon Dec 05 17:16:06 UTC 2016
Pages17
Paragraphs1026
Words3214

Comments