Note that derby uses 64 bit file addressing, so as jack says it is limited by the files system support up to 64bit addressing.
If you need a database that is bigger than a single disk, if your OS supports it you need to map multiple disks so that to the JVM it looks like a single disk under the directory. Derby provides no spreading of a single table across disks (either striping or union). The assumption is that the OS/filesystem/disk controller can do this below derby. So the derby architecture has few upper limits on size of data, but it does not have features one might expect in a VLDB (ie. parallel thread execution of a single query, parallel sort, offline indexing, incremental backup, ...). Jack Klebanoff wrote: > Hong Ji wrote: > >> Hi, >> >> Does Derby have any limit on the database size? >> >> >> >> Thanks. >> >> Hong >> > A Derby database is stored in multiple files under one directory, so the > database size is limited by the file system. The maximum database size > is limited by the maximum that the file system allows in one partition. > It is system dependent. > > By default the log files are stored on the same partition as the data, > further limiting the database size. However the log files can be stored > on a separate partition. > > Jack >
