chunk transplant
Posted in 2001
> I thought Robert said his server OS mirrored the chunks, so if the > hardware fail, he has other server with the same configuration and > hardware able to bring on-line a new Ifx server by using the mirrored > chunks. It works, but if the IDS get a unrecoverable crash, this schema > wont work. > But mirror is not necessary; you could take the disks to the other server > or even better, you could have the disks, where the chunks are, connected > to both servers, but activated only in one of course (you would need extra > software). > > -----Mensaje original----- > De: David Williams [SMTP:djw@smooth1.demon.co.uk] > Enviado el: Martes 16 de Enero de 2001 07:31 PM > Para: informix-list@iiug.org > Asunto: Re: chunk transplant > > In article <9422kf$k08$1@knight.vf.lmco.com>, Robert Carts > <rcarts@agccs.lmco.com> writes > >We have been investigating ways to keep data current on a local and > >identical backup server. We are aware of IER and HDR. My boss > suggested > >that we keep a mirrored chunk on a cross mounted drive that resides > on the > >backup server. Then in the event of a failure of the primary > server, bring > >the backup server online using the primary servers chunk. I > skeptically > >attempted this and it worked! > > > >The conditions are: The schemas, database spaces, chunk size and > number of > >chunks are the same on both servers. The chunks use cooked files. > The > >databases are created in rootdbs and the tables, indexes are > created in > >dedicated chunks. The chunks from the primary server are copied > over the > >backup server chunk files before the backup instance is brought > online. We > > ?? If the primary server is down how can you do the copy? > > I'm not sure I follow, take the root dbspace as an example. What > chunks are on each machine. What is the copying which is done? > > > >are using 9.2uc3 on solaris 2.7. > > > >Question: Under what conditions would this not work (within the > conditions > >above) ? I either need to demonstrate why not to use the technique > or make > >it work reliably? > > > >Thanks, Bob Carts > > > > > > -- > David Williams