Re: [Info-Ingres] Recover checkpoint to server withdifferent locations -
11-30-2011
, 04:39 AM
Hi Kyle,
I've never tried the relocatedb style you mentioned...why not give it a try and let us know how you get on.
But it sounds like the symlink option on your DR box is viable. All you need is to get root to do:
mkdir /Ingres
cd /Ingres
ln -s /opt/IngresII ing_prod
Afterwards on your DR box any reference to the production path /Ingres/ingres_prod/ingres/data/default/dbname will actually traverse the equivalent DR path /opt/IngresII/ingres/data/default/dbname
You need to ensure that every location that is used on your prod box has an identical name on your DR box.
So if your prod box has data location names ii_database, ii_database_2, ii_database_3...then these names also exist on your DR box. Ditto for checkpoint, journal, dump work and auxilliary work locations.
Now just create the empty database on the DR host, using the same location names as on the prod host. Checkpoint the database and turn journaling on (assuming its on in the prod version).
You should now be in a position to copy over the config files from the prod host....as a tip ALWAYS save the original DR versions aside before this. Ditto copy in the checkpoint, journal and dump files. The latter include [dc]*.dmp...you need these!
Do an infodb ... check it now looks like the prod version. Check each path mentioned in the output can be accessed.
Do the rollforwarddb
Do alterdb -u... dbname -disable_journaling
then ckpdb +j dbname.
These last two steps are necessary to get the log file on the DR host to play ball. Otherwise you'll find the recovered database marked inconsistent real soon. If the database is too large to want to do a full ckpdb +j, I can show you how to monkey with a cktmpl.def to allow it to do this step without actually taking a backup...just doing the '+j' bit.
Marty
-----Original Message-----
From: Ingres Forums [mailto:info-ingres (AT) kettleriverconsulting (DOT) com]
Sent: 29 November 2011 12:55
To: info-ingres (AT) kettleriverconsulting (DOT) com
Subject: Re: [Info-Ingres] Recover checkpoint to server with different locations
Thanks everyone for your replies. It helps confirm the situation that
I figured I'm in, of which is changing locations within a checkpoint is
not really meant to be.
Is there any chance using 'relocatedb --location --newlocation' type
command could help. Here's the overview of what I'm thinking:
- I have a checkpoint from production server I'd like to
recover(it has say 20 data locations all mapped high in folder tree
under /INGRES/ing_prod)
- On the DR box, I've created an empty DB of which I want to
recover this checkpoint.
- This DR box also has about 20 locations but they're all mapped
with a different high folder path of: /opt/IngresII
- Pull over the aaaaaaaa.cnf from the production box to the DR
box
- setup 20 new data locations on DR box to match folder paths
within aaaaaaaa.cnf from production
- use relocatedb -location= -newlocation= to change
aaaaaaaa.cnf to use new locations
I haven't any experience with relocatedb, so I'm hoping it adjusts the
aaaaaaaa.cnf file locations, so maybe your thoughts on this and the
above scenario if it'd get me close to working. I guess if the above
could work, the last laundry item I have is the default data location
because these are different folder paths as well.
New installation definitely sounds the way to go... Thanks again for
your help.
--
kylehanson
_______________________________________________
Info-Ingres mailing list
Info-Ingres (AT) kettleriverconsulting (DOT) com
http://ext-cando.kettleriverconsulti...fo/info-ingres |