Re: Replication on the same box ????

In article <54nd5v$3tn@nntp.idgonline.no>, Nils Myklebust <Nils.Myklebust@idg.no> writes >David Williams <djw@smooth1.demon.co.uk> wrote: > >: Sometime I wish I would experiment rather than reading the manuals and >: thinking I understand them. I just tried (~6:30pm today) to initailise >: two Online Instances with ROOTPATH set to ./root1 in separate >: directories. It worked! When the Informix manuals say full pathname >: they just mean pathname.. you can use relative pathnames!!. > >: I managed to get a level-0 archive done with tape devices set to >: ./tapedev and ./archdev. But I didn't get time to finish setting up >: the data replication. I'll finish it tomorrow morning/lunchtime and >: I'll post my findings tomorrow. > >: Don't rule out data replication on the same machine quite >: yet...possibly an extra line in the Informix manuals about relative >: pathnames are Ok might be in order...I'll report back tomorrow... > >I haven't seen your report back yet, but you might want to be carefull >with relative pathnames in a production environement. One might guess >that Informix would use the pathnames relative to the current >directory when you start up the engine. If someone forgets one day >(two years later perhaps) it might create havock. There may be other >problems as well. Of course for testing purposes it is ok, but there >is a good reason why the Informix manuals says to use full pathnames. >That's the only way to ensure the engine will allways start correctly >using the correct devices. > >If you can get replication to run on one machine it would be a good >thing for everyone to be able to test out how it works without having >to have two machines. But there is no reason to run that way in >production, is there? > > >Nils.Myklebust@idg.no >NM Data AS, P.O.Box 9090 Gronland, N-0133 Oslo, Norway >My opinions are those of my company > I got it too work, hurrah!. As others have said you need to be very careful. I'm only using it to play around with data replication so I can have some fun and put it on my CV. I would NEVER do this in a live enviroment....but it is fun. I had a few problems due to trying this you really shouldn't do so I may have some questions soon Eg. I had it set to have the secondary automatically become the primary when the primary failed. I then shutdown the primary and the seconday took over. BUT when I restarted the primary starnge things happen... eventually the one which had taken over as the primary ended up with a checkpoint record in two different logical logs (onstat -l showed a 'C' on 2 adjacent logical logs!!). I think some more experimentation is in order and I'll keep better notes on exactly what I do... -- David Williams