Re: No ODBC access to Informix...
Posted in 2004
On Mon, 08 Nov 2004 09:36:02 -0500, Wolf wrote: You could use the HDR (High availability Data Replication) of IDS to create a read-only backup server instance on another machine (in theory it could be on the same machine, but I wouldn't try it). HDR replicates the entire server instance on the backup which is read-only while the primary is online. If the primary goes offline users reconnect to the back server which is manually switched to stand-alone mode (ie read-write not replicated) until the primary comes back online. At that time one can either make the original primary secondary or if the users will tolerate it, bring the primary back up-to-date and make it primary again. For your purposes you would only publish the backup server for ODBC connections. Art S. Kagel > Hello all, > > after googling I saw that the same discussion came up in 1997/98 - but did > not lead a satisfactory result for me.... > > Situation: > We have some kind of a pre-fab ERP solution in house. From any PC a telnet > session is used to connect to the Solaris and to start the application. > > The application uses the Solaris user ID to identify itself to the database > - and hence there are certain rights, that those users have. E.g. a user may > well have the right to change a record using the application interface's > business logic - BUT may not have the right to change a row arbitrarily. > > So far reporting has been done using MS Query/ODBC. Of course any user could > simply go in MS Query and generate any SQL code - basically allowing him/her > to circumvent the application's business logic.... > > I do not know Informix - but from my understanding there must be a way to > totally disable ODBC access towards this database. I know a little about DB2 > and Oracle and from there I have the following picture: > > - every box may have one or more instances of a database "engine" - every > engine may manage one or more databases > > Is it like this with Informix (5.01)? Is there a "listener" per engine, that > can be tweaked the way I want it? > > We want to create a "report" DB - either on a seperate machine or - > preferably - on the same machine, where MS Query access would be allowed.... > > So my ideal picture looks like this: > - one Solaris box > - two instances of Informix loaded > --- one w/o ODBC access > --- one w/ ODBC access > - every instance managing one DB > - applying the "redo"-logs of the w/o ODBC engine towards the other DB gives > a time-shifted "copy" of the production DB. > > Does this make sense at all????????????? > > Any postings appreciated! > > Regards > Wolf Duttlinger-Manger