Results 1 to 5 of 5
  1. #1
    Join Date
    Nov 2006

    Question Database recovery!

    ok before you ask this is a homework question, but i just need some guidance and want to see if im on the right track.

    first the scenario.
    im a DBA responsible for a DB that performs on line transaction processing. no mirroring technique is being used due to cost. its 14:30 on a friday a disk head crash occurs.
    out line the steps that u would take BEFORE and AFTER the event.

    ok for before, you need to make sure you have a working backup. once a working backup has been restored, after the event, a rollback would be performed to just before the crash occured.

    just want to know if im on the right thinking track? thanks for any help.

  2. #2
    Join Date
    Jun 2003
    You are on the right track. Now what's a "rollback", how would you perform it, and what is required for it to be successful?
    If it's not practically useful, then it's practically useless.

    blindman "sqlblindman"

  3. #3
    Join Date
    Nov 2004
    out on a limb
    so you've got a cost problem... disk mirroring isnt that expensive, what about RAID, presumably server clustering is also out the window.

    your organisation has got an operational problem.. how significant is the loss of that disk to the operation of the company (ie how soon MUST they have that system and whatever data is on it back

    you have a disaster planning problem (do you have spare capacity elsewhere on the network, have you chose good locations for critical elements of your data [there is stuff all point having a backup on the same device.. I ha\ve seen servers with a tape drive attached to the same server as critical data, Ive seen the ldb log files on the same physical disk as the main db]).

    you have a transient data problem how to recover any data between the lat backup and the time the system crashed

    you may have a data integrity problem if your developers dont properly design the applciation (ie all elements of the transaction stand or fail together)... but thats not really a specific issue with a disk headbut.. thats an issue of competent design, that should cater for hardware failures, but also other data or software contingencies
    I'd rather be riding on the Tiger 800 or the Norton

  4. #4
    Join Date
    Nov 2006
    thanks for the replies. the loss of the disk is critical as the company requires up to the minute recovery. also it says details should be included of what help the database management system itself would give me. is that referring to things like the redo log files and rollbacks?

  5. #5
    Join Date
    Jan 2003
    After the event, I would seriously consider asking the lead accountant how much money the outage cost the company, and compare that to the cost of a properly redundant system.

    I assume you are working with Oracle, since you are talking about redo logs?
    The only problem I actually have with your solution so far is the word "rollback".

    As to what the RDBMS does for you, consider what conditions need to be satisfied before a database can be opened. Actually, it may be hard for us to guess what the professor wants in this question, so that may have to be taken with a largish grain of salt.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts