Manually rebuilding Adaptive Server with bcp and server building utilities enables you to create a new master device/configuration block and preserve system tables.
As of version 12.5, master devices and servers are built with the dataserver (Unix) and sqlsrvr (Windows) programs. In pre-12.5 versions, they are built with buildmaster.
Some of the most common uses for a manual rebuild are when:
The master device has no more available space. You can migrate the system table information to a new, larger master device while retaining all current data on the original devices.
Restoring the master device resulted in 605 errors due to an incorrect sysusages table. You can use trace flags, along with the bcp and server build steps, to get the information needed for a rebuild.
No backup of the master database exists. The old master database is accessible, although it is not runnable. You can migrate the system information to a new master database.
You can also use manual rebuild steps for immediate recovery when:
Severe data corruption necessitates a speedy recovery. You can run the bcp and server build commands, instead of using the sybinit utility, running the disk reinit/refit commands, or creating/loading from backups.
You need to perform a recovery from an inadvertent configuration change, such as memory set too high.
Major corruption problems on, or loss of, the master device requires creating a new server. In this case, you must use bcp on your system tables immediately.
Ensure that your backup procedures include bcp commands
for all relevant tables, including the six system tables listed
under “Steps for Rebuilding Adaptive Server”.
You can then more easily restore the master database
if necessary.
