Marcin,

> I wonder if shadow file would be a good option to have working copy of

the

> database. I will have two linux server connected through 1G LAN. As my
> linux guru said - there is no problem to prepare some kind of cluster,

so

> the shadow file would be phisically written to thed disk located in

the

> second server. In case of main server failure we could easily switch

to

> the second server having only a hlaf hour brake.

Shadowing has created before economical hardware based disk
mirroring/RAID drive solutions became available. While shadowing can be
used, it's performance will be VERY POOR when compared to hardware or OS
software mirroring solutions.

What you are looking for is database replication, or improving the
configuration of your hardware/disk redundancy (you could configure a
backup server which could be up and running within minutes)


Sean