home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!zaphod.mps.ohio-state.edu!sample.eng.ohio-state.edu!purdue!bu.edu!decwrl!decwrl!ogicse!emory!mnopltd!neal
- From: neal@mnopltd.UUCP
- Newsgroups: comp.databases
- Subject: Re: Simple Warm Standby DBMS's -
- Message-ID: <252@mnopltd.UUCP>
- Date: 7 Sep 92 20:13:47 GMT
- Reply-To: emory!mnopltd!neal
- Organization: MNOP Ltd., Atlanta, Ga.
- Lines: 29
-
- There has been a lot of interesting discussion on this problem. Here is
- a simple arrangement which does seem to work:
-
- Progress RDBMS Version 5 and higher support a separate after-image file.
- This file can be forced over to an RFS mounted file system on the hot system.
-
- After failure, the roll-forward will bring the database up to the last
- complete transaction written. Thus, recovery time is about 20 minutes,
- depending on how long since a backup copy of the primary db was placed on
- the secondary.
-
- Version 7 may avoid doing seeks on the after image file, so it could be
- placed on /dev/rct0, thus avoiding the networking. Use on the spare
- involves pulling out tape, inserting, doing roll forward.
-
- If you can stand having the system down for a few minutes several times
- a day the AI files can be switched, and you can run the roll forward on
- the spare, keeping it closer to the contents of the live system and
- making failure cutover quicker.
-
- Now I wouldn't use this for Air Traffic Control, but if you can afford
- to be down for 20 minutes it might work.
-
-
- ------------------------------------------------------------------------------
- Neal Rhodes MNOP Ltd (404)- 972-5430
- President Lilburn (atlanta) GA 30247 Fax: 978-4741
- emory!mnopltd!neal
- ------------------------------------------------------------------------------
-