Data Consistency, Recovery, and Migration 4-15
Beginning
Recover a mirrored chunk through the DB-Monitor Dbspaces menu, Status
option or through the tbspaces utility (change the status of the failed chunk
from down, D, to online, O). When you initiate recovery, OnLine puts the
down chunk in recovery mode and copies the information from the online
chunk to the recovery chunk. When the recovery is complete, the chunk
automatically receives online status. You perform the same steps whether
you are recovering the primary chunk of a mirrored pair or recovering the
mirror chunk. (For information about monitoring chunk status codes, refer to
page 3-70.)
Beginning
Mirroring begins immediately after you create the mirror chunk, in most
cases. (The one exception occurs when you start mirroring for a dbspace that
contains logical log files. This topic is addressed in the next section.)
The new mirror chunk enters recovery mode automatically. A mirror-recovery
process copies information from the primary chunk to the mirror chunk.
When the two chunks are considered equal, the mirror chunk automatically
receives online status (O). From this point on, the mirror chunk reports that it
has zero free pages.
Data from the online chunk is copied to the chunk in recovery in eight-page
increments. During the copy, these blocks of pages are frozen. Any modifica-
tions that affect those pages must wait until the copy to the recovery chunk
is complete.
Recovery time is almost instantaneous when the primary chunk and mirror
chunk are being created in the same operation. If you are adding a mirror
chunk to an existing chunk, the recovery time depends on the amount of data
in the chunk.
The recovery procedure that marks the beginning of mirroring is delayed if
you are starting to mirror a dbspace that contains a logical log. Mirroring for
this dbspace does not begin until you create a level-0 archive.