apply TX logs after data migration

We are planning a migration from from one platform to another (both Intel/Redhat, IDS 10.00.UC4, new one is bigger, faster, stronger, etc....). We want to do the migration from a staging server to: - leave production up - move only 1 years worth of data; leave historical data behind The current plan is to restore a level-0 archive to the staging server and then use HPL to unload from it and load into the new configuration. We'll change the TX logging tape about the same time we start the level-0. When we get the new system ready, the current system will be stopped and we'll pull the TX log tape. The questions: - will the logical restore (ontape -l) get us current with production? - will there be an issue if the new configuration has tables in 8K dbspaces and the current system is mostly 2K spaces with some 8K spaces? Anyone with experience doing this? TIA. Bob Roussey

On 15/12/06, ROBERT ROUSSEY <robert.roussey@spiritair.com> wrote: > > We are planning a migration from from one platform to another (both > Intel/Redhat, IDS 10.00.UC4, new one is bigger, faster, stronger, etc....). We > want to do the migration from a staging server to: > - leave production up > - move only 1 years worth of data; leave historical data behind > > The current plan is to restore a level-0 archive to the staging server and > then use HPL to unload from it and load into the new configuration. We'll > change the TX logging tape about the same time we start the level-0. When we > get the new system ready, the current system will be stopped and we'll pull > the TX log tape. > The questions: > - will the logical restore (ontape -l) get us current with production? > - will there be an issue if the new configuration has tables in 8K dbspaces > and the current system is mostly 2K spaces with some 8K spaces? > > Anyone with experience doing this? > > TIA. > > Bob Roussey > > > ******************************************************************************* > Forum Note: Use "Reply" to post a response in the discussion forum. > > Bob If I understand correctly, you want to restore level 0 to an intermediate server, copy selected data to your final server and then apply the log to this final server. This is not possible. The log can only be applied to the intermediate server as part of the restore. You will need to freeze your production system, or do something creative with ER in order to achive this type of cutover. Using HPL from 2 K pages on the intermediate server to 8 K on the final server is not a problem. Keith

Hi, I'm not 100% clear what your plan is. The part with HPL troubles me ... I gather you want to use HPL to load data into dbspaces with different pagesize (and maybe some other config changes in the new system). This sounds reasonable. However, if you do this, the new instance (new system) will be in On-Line mode for the load of all this data. In this mode you cannot restore logical log files. Restore of logical log files is only possible during the logical log restore phase. And you get into this phase only during a restore (from an archive). In your case you would start the new instance from scratch and then load the data via HPL. Therefore this instance would not be restored from an archive. And therefore you cannot restore logical log files to such an instance. If you restore the new instance directly from the level-0 archive, then you would also be able to restore the logical log file that was backed up on the other instance after the archive was done. But that way you would not have the configuration changes (like different page size for dbspaces). Regards, Martin -- Martin Fuerderer IBM Informix Development Munich, Germany Information Management Informix URLs list: http://home.arcor.de/mfu1/informix/urls.html ids-bounces@iiug.org wrote on 15.12.2006 20:49:11: > > We are planning a migration from from one platform to another (both > Intel/Redhat, IDS 10.00.UC4, new one is bigger, faster, stronger, etc....). We > want to do the migration from a staging server to: > - leave production up > - move only 1 years worth of data; leave historical data behind > > The current plan is to restore a level-0 archive to the staging server and > then use HPL to unload from it and load into the new configuration. We'll > change the TX logging tape about the same time we start the level-0. When we > get the new system ready, the current system will be stopped and we'll pull > the TX log tape. > The questions: > - will the logical restore (ontape -l) get us current with production? > - will there be an issue if the new configuration has tables in 8K dbspaces > and the current system is mostly 2K spaces with some 8K spaces? > > Anyone with experience doing this? > > TIA. > > Bob Roussey > > > ******************************************************************************* > Forum Note: Use "Reply" to post a response in the discussion forum. >

HI, Bob, A option is, (1) restore the leve 0 on another plate form (2) prepare a script to load(copy) all newly genertaed data( after level 0) to the new instance( you definitely only need copy the data table you want , NOT whole database tables, like temporary, asistant, ...) (3) stop production , run copy script( do it online) (4) start both new and production dbs, now they are fully synchronized. Frank On 12/15/06, ROBERT ROUSSEY <robert.roussey@spiritair.com> wrote: > > > We are planning a migration from from one platform to another (both > Intel/Redhat, IDS 10.00.UC4, new one is bigger, faster, stronger, > etc....). We > want to do the migration from a staging server to: > - leave production up > - move only 1 years worth of data; leave historical data behind > > The current plan is to restore a level-0 archive to the staging server and > then use HPL to unload from it and load into the new configuration. We'll > change the TX logging tape about the same time we start the level-0. When > we > get the new system ready, the current system will be stopped and we'll > pull > the TX log tape. > The questions: > - will the logical restore (ontape -l) get us current with production? > - will there be an issue if the new configuration has tables in 8K > dbspaces > and the current system is mostly 2K spaces with some 8K spaces? > > Anyone with experience doing this? > > TIA. > > Bob Roussey > > > > ******************************************************************************* > Forum Note: Use "Reply" to post a response in the discussion forum. > >