Vision acquires Lakeview (Mimix) - IBM AS400

This is a discussion on Vision acquires Lakeview (Mimix) - IBM AS400 ; http://www.lakeviewtech.com/pressroom/pr_2007_06_13.asp Also http://new.marketwire.com/2.0/release.do?id=741691...

+ Reply to Thread
Results 1 to 8 of 8

Thread: Vision acquires Lakeview (Mimix)

  1. Vision acquires Lakeview (Mimix)


  2. Re: Vision acquires Lakeview (Mimix)

    On Jun 13, 10:23 am, Jonathan Ball wrote:
    > http://www.lakeviewtech.com/pressroom/pr_2007_06_13.asp
    >
    > Alsohttp://new.marketwire.com/2.0/release.do?id=741691


    what does it take to write high availability software for the as400?
    Is it just a matter of capturing journal entries and applying them to
    the database of another system?

    -Steve


  3. Re: Vision acquires Lakeview (Mimix)

    On Jun 13, 9:28 am, Steve Richter wrote:
    > On Jun 13, 10:23 am, Jonathan Ball wrote:
    >
    > >http://www.lakeviewtech.com/pressroom/pr_2007_06_13.asp

    >
    > > Alsohttp://new.marketwire.com/2.0/release.do?id=741691

    >
    > what does it take to write high availability software for the as400?
    > Is it just a matter of capturing journal entries and applying them to
    > the database of another system?


    I think there's a little more to it than that, but that's the gist of
    the database replication. However, the HA packages will detect
    changes to objects and synch up the entire object as well. They're
    fairly sophisticated packages.


  4. Re: Vision acquires Lakeview (Mimix)

    On Jun 13, 12:33 pm, Jonathan Ball wrote:
    > On Jun 13, 9:28 am, Steve Richter wrote:
    >
    > > On Jun 13, 10:23 am, Jonathan Ball wrote:

    >
    > > >http://www.lakeviewtech.com/pressroom/pr_2007_06_13.asp

    >
    > > > Alsohttp://new.marketwire.com/2.0/release.do?id=741691

    >
    > > what does it take to write high availability software for the as400?
    > > Is it just a matter of capturing journal entries and applying them to
    > > the database of another system?

    >
    > I think there's a little more to it than that, but that's the gist of
    > the database replication. However, the HA packages will detect
    > changes to objects and synch up the entire object as well.


    I see. how would that be done? Using the audit journal and system
    exit programs, I guess. Curious to know if there are any unpublished
    APIs that the HA software companies pay to have access to that an open
    source project would not be able to use.

    -Steve




  5. Re: Vision acquires Lakeview (Mimix)

    il 13/06/2007 18.28, Scrive Steve Richter 40341048:
    > On Jun 13, 10:23 am, Jonathan Ball wrote:
    >> http://www.lakeviewtech.com/pressroom/pr_2007_06_13.asp
    >>
    >> Alsohttp://new.marketwire.com/2.0/release.do?id=741691

    >
    > what does it take to write high availability software for the as400?
    > Is it just a matter of capturing journal entries and applying them to
    > the database of another system?
    >
    > -Steve
    >

    That's not High Availability, it's just data-replication.
    It's not the same.

    --
    Dr.Ugo Gagliardelli,Modena,ItalyCertifiedUindoscrasherAñe joAlcoolInside
    Spaccamaroni andate a cagare/Spammers not welcome/Spammers vão à merda
    Spamers iros a la mierda/Spamers allez vous faire foutre/Spammers loop
    schijten/Spammers macht Euch vom Acker/Spamerzy wypierdalac'

  6. Re: Vision acquires Lakeview (Mimix)

    Beyond audit journal, the journal of the journaled objects; the
    journal entry APIs. For the database objects, beyond data auditing,
    refer to the API:
    http://publib.boulder.ibm.com/infoce...s/qdbrplay.htm

    Note, a tidbit, if embarking on the alluded project: A common problem
    over the years for HA sfw had been overzealous attempts to perform some
    action on the object ASAP after T-CO ['Object created' audit entry],
    e.g. a SAVOBJ, such that an application doing some action soon after the
    create is then locked out by that [save] activity. Either the
    application must be aware of the possibility of the conflict or the HA
    sfw needs to temper its rush to accommodate the typical actions of an
    application; e.g. GRTxxx, ADDmbr, CHGxxx, etc. which might be performed
    by the app _after_ the create activity to 'finish creating' the object
    as the application intends/requires.

    Regards, Chuck
    -- All comments provided "as is" with no warranties of any kind
    whatsoever and may not represent positions, strategies, nor views of my
    employer

    Steve Richter wrote:
    > On Jun 13, 12:33 pm, Jonathan Ball wrote:
    >> On Jun 13, 9:28 am, Steve Richter wrote:
    >>
    >>> On Jun 13, 10:23 am, Jonathan Ball wrote:
    >>>> http://www.lakeviewtech.com/pressroom/pr_2007_06_13.asp
    >>>> Also http://new.marketwire.com/2.0/release.do?id=741691
    >>> what does it take to write high availability software for the as400?
    >>> Is it just a matter of capturing journal entries and applying them to
    >>> the database of another system?

    >> I think there's a little more to it than that, but that's the gist of
    >> the database replication. However, the HA packages will detect
    >> changes to objects and synch up the entire object as well.

    >
    > I see. how would that be done? Using the audit journal and system
    > exit programs, I guess. Curious to know if there are any unpublished
    > APIs that the HA software companies pay to have access to that an open
    > source project would not be able to use.
    >
    > -Steve


  7. Re: Vision acquires Lakeview (Mimix)

    On Jun 13, 2:59 pm, CRPence wrote:
    > Beyond audit journal, the journal of the journaled objects; the
    > journal entry APIs. For the database objects, beyond data auditing,
    > refer to the API:http://publib.boulder.ibm.com/infoce...ndex.jsp?topic...


    interesting. did not know about that api.

    > Note, a tidbit, if embarking on the alluded project: A common problem
    > over the years for HA sfw had been overzealous attempts to perform some
    > action on the object ASAP after T-CO ['Object created' audit entry],
    > e.g. a SAVOBJ, such that an application doing some action soon after the
    > create is then locked out by that [save] activity. Either the
    > application must be aware of the possibility of the conflict or the HA
    > sfw needs to temper its rush to accommodate the typical actions of an
    > application; e.g. GRTxxx, ADDmbr, CHGxxx, etc. which might be performed
    > by the app _after_ the create activity to 'finish creating' the object
    > as the application intends/requires.


    I dont follow. The replicate - from system does the following:
    - job a crtpf
    - job a grtobjaut
    - job b savobj
    - job a chgpfm
    - job a insert some rows

    all these transactions will be in the journal and can be applied
    sequentially to the replicate-to system, no?



  8. Re: Vision acquires Lakeview (Mimix)

    That comment only applies to _if_ the implementation of the
    replication were to use SAVOBJ to generate the first copy of the file at
    the source system to create/restore at the target system. Using QDFTJRN
    and replaying the create for the journaled file should suffice.

    An example of failures if/when SAVOBJ is used to replicate:
    RUNQRY *NONE QIWS/QCUSTCDT OUTFILE(HATRACKED/MYCOPY) AUT(*EXCLUDE)
    for example would generate a new file without CRTPF nor CREATE TABLE.
    With Audit entries the T-CO identifies the file, but there is no
    information about its somewhat spontaneous creation. So... If
    implementation of replication for the T-CO were by SAVOBJ to send the
    file to the other system, then that save could interfere with a later
    request by the application to GRTOBJAUT HATRACKED/MYCOPY *FILE CRP
    AUT(*ALL)
    Thus 'job b' doing SAVOBJ between create and grant can break the
    application running in 'job a' because it does not expect anyone to be
    locking the file that was just created. Of course the application
    should be coded to deal with that; for me, that would be a CLP without
    any MONMSG ensuring that the program would go into inquiry wait on CPA0701.

    Note that was just an example: I do not advocate having ad-hoc query
    outfiles replicated [or used in production] -- use CRTPF or CREATE TABLE
    instead.

    Regards, Chuck
    -- All comments provided "as is" with no warranties of any kind
    whatsoever and may not represent positions, strategies, nor views of my
    employer

    Steve Richter wrote:
    > On Jun 13, 2:59 pm, CRPence wrote:
    >> Beyond audit journal, the journal of the journaled objects; the
    >> journal entry APIs. For the database objects, beyond data auditing,
    >> refer to the API:http://publib.boulder.ibm.com/infoce...ndex.jsp?topic...

    >
    > interesting. did not know about that api.
    >
    >> Note, a tidbit, if embarking on the alluded project: A common problem
    >> over the years for HA sfw had been overzealous attempts to perform some
    >> action on the object ASAP after T-CO ['Object created' audit entry],
    >> e.g. a SAVOBJ, such that an application doing some action soon after the
    >> create is then locked out by that [save] activity. Either the
    >> application must be aware of the possibility of the conflict or the HA
    >> sfw needs to temper its rush to accommodate the typical actions of an
    >> application; e.g. GRTxxx, ADDmbr, CHGxxx, etc. which might be performed
    >> by the app _after_ the create activity to 'finish creating' the object
    >> as the application intends/requires.

    >
    > I dont follow. The replicate - from system does the following:
    > - job a crtpf
    > - job a grtobjaut
    > - job b savobj
    > - job a chgpfm
    > - job a insert some rows
    >
    > all these transactions will be in the journal and can be applied
    > sequentially to the replicate-to system, no?


+ Reply to Thread