Cloning Production to Dev/Test/Train/Regression

Hi All, Not sure if I'm after a new feature or just some scripts, surely it's what us DBA's do. I'm just wondering the easiest (laziest as I'm busy) way to copy our production database to say development, training, test etc. once we upgrade to V11.5. With all of the new CDR features in V10 upwards, is there some way I can say - setup database and sync. Then at a later point reconcile (data or everything) whilst dropping CDR in-between to negate any CDR risks. I guess we could go down the HDR path, but then you can run the same potential problems with networks, transactions. Also we currently use logged extended datatypes - spatial, web, etx which are always a headache with spatial clustering etc.. The only thing you run into generally is that you have different user permissions across different environments (developers different to production). Also synonyms / views sometimes vary across environments pointing to different non production sources so you wouldn't always want the entire schema or user permissions across. Also, is there a way to rename chunks in a SAN clone operation yet without dd'ing & patching the reserved pages thus negating the need for a slow and potentially error prone external rename chunk restore? This SAN clone option might be quicker then just adjust any views / synonyms / permissions. Yes, I know about dbexport/onpload/hpl etc.. just wondering if there is a really simple easy foolproof fast way with all the new V10 features. Any ideas appreciated. Its late in OZ, so excuse crap grammer etc.. TIA, Peter.

PeterP wrote: > Hi All, > > Not sure if I'm after a new feature or just some scripts, surely it's > what us DBA's do. > > I'm just wondering the easiest (laziest as I'm busy) way to copy our > production database to say development, training, test etc. once we > upgrade to V11.5. backup and restore to a different machine? > > With all of the new CDR features in V10 upwards, is there some way I > can say - setup database and sync. Then at a later point reconcile > (data or everything) whilst dropping CDR in-between to negate any CDR > risks. > I guess we could go down the HDR path, but then you can run the same > potential problems with networks, transactions. Also we currently use > logged extended datatypes - spatial, web, etx which are always a > headache with spatial clustering etc.. > > The only thing you run into generally is that you have different user > permissions across different environments (developers different to > production). Also synonyms / views sometimes vary across environments > pointing to different non production sources so you wouldn't always > want the entire schema or user permissions across. > > Also, is there a way to rename chunks in a SAN clone operation yet > without dd'ing & patching the reserved pages thus negating the need > for a slow and potentially error prone external rename chunk restore? > This SAN clone option might be quicker then just adjust any views / > synonyms / permissions. > > Yes, I know about dbexport/onpload/hpl etc.. just wondering if there > is a really simple easy foolproof fast way with all the new V10 > features. Any ideas appreciated. > > Its late in OZ, so excuse crap grammer etc.. > TIA, Peter.

On Mar 4, 2:14 am, Madison Pruet <mpru...@verizon.net> wrote: > PeterP wrote: > > Hi All, > > > Not sure if I'm after a new feature or just some scripts, surely it's > > what us DBA's do. > > > I'm just wondering the easiest (laziest as I'm busy) way to copy our > > production database to say development, training, test etc. once we > > upgrade to V11.5. > > backup and restore to a different machine? > > > > > With all of the new CDR features in V10 upwards, is there some way I > > can say - setup database and sync. Then at a later point reconcile > > (data or everything) whilst dropping CDR in-between to negate any CDR > > risks. > > I guess we could go down the HDR path, but then you can run the same > > potential problems with networks, transactions. Also we currently use > > logged extended datatypes - spatial, web, etx which are always a > > headache with spatial clustering etc.. > > > The only thing you run into generally is that you have different user > > permissions across different environments (developers different to > > production). Also synonyms / views sometimes vary across environments > > pointing to different non production sources so you wouldn't always > > want the entire schema or user permissions across. > > > Also, is there a way to rename chunks in a SAN clone operation yet > > without dd'ing & patching the reserved pages thus negating the need > > for a slow and potentially error prone external rename chunk restore? > > This SAN clone option might be quicker then just adjust any views / > > synonyms / permissions. > > > Yes, I know about dbexport/onpload/hpl etc.. just wondering if there > > is a really simple easy foolproof fast way with all the new V10 > > features. Any ideas appreciated. > > > Its late in OZ, so excuse crap grammer etc.. > > TIA, Peter. Thanks Madison, that's not a bad choice, was after something easier/ more ER sync like which I've not had much time to read about. I did mention but should have made it clearer: 'thus negating the need for a slow and potentially error prone external rename chunk restore?' Also I was wondering on options to copy data from say 98% of the tables and reconcile now and again. And then there's keeping the spatial clustering in sync ... sigh ... Vnext auto extent merge/ cluster/index reorg maybe?