HP NonStop Shadowbase Command Definitions
Collector Command Descriptions
HP NonStop Shadowbase Command Reference Manual—785421-002
96
NUM ABORTING shows the number of aborted transactions currently being
tracked.
ABORT TIMER shows whether the ABORT TIMER is on or off for the aborted
transactions currently being tracked. (The ABORT TIMER helps resolve
certain specific TMF recovery scenarios such as when a TMF
failure/restart has occurred).
NUM PEER TRANS NO AUDIT shows the number of the peer-CONS’s pre-started
transactions that have not yet yielded audit events.
RESERVED BUFFER INFO – This section displays reserved buffer
statistics. These buffers are allocated and used to pre-read the audit trail
data in preparation for sending them to the Consumer(s). If the Collector
is not "caught up" (i.e.,i.e., the number of EOFs is not increasing), one can
attain higher throughput if there are reserved buffers in use (The reserved
buffers are ready and waiting for the Consumer(s) to be available to take
their contents).
NUM ALLOC shows the number of reserved buffers allocated.
CUR USED shows the number of reserved buffers currently in use.
MAX USED shows the maximum number of reserved buffers that have been
in use since the STATS started to be gathered.
BI-DIRECTIONAL PEER CONSUMER FAILURE INFO – This section
displays the timestamp of the last peer consumer failure and the total
number of times it has failed. This section is only displayed for bi-
directional environments and only when the peer consumer has
failed at least once.
VOLUME/FILE RECOVERY INFO – This section displays information
about any volume or file recovery that occurred for this collector. It shows
the number of times recovery happened as well as the timestamp and
audit trail location of the first and last recovery events. This section is
only displayed after the first time volume or file recovery occurs for
the collector.
PRESCREEN SUBVOLUMES – The Collector will build a list of up to 8
distinct subvolume names (not including the volume name, like $DATA1)
from the SOURCEFILE parameter for all of your DBS objects. This list will
be checked first for a match against the file name for an incoming audit
trail event. If a match doesn’t occur, the Collector will discard the event
immediately and continue with the next event in the audit trail (this
“prescreening” is occurring sooner in the selection cycle than in the past).
This approach will optimize the Collector’s cpu usage for environments
where there is lots of activity for many different subvolumes on a system
and the related Collector is not selecting these subvolumes for replication.
For the second part, called Filename Prescreening (recorded under the
heading
‘FILE NAME SEARCH LIST INFO’), as well as for environments where
Komentarze do niniejszej Instrukcji