This table provides information regarding the latency of disk operations for NDB Cluster Disk Data tablespaces.
The following table provides information about the columns in
pgman_time_track_stats table. For each
column, the table shows the name, data type, and a brief
description. Additional information can be found in the notes
following the table.
Table 22.322 Columns of the pgman_time_track_stats table
||integer||Unique node ID of this node in the cluster|
||integer||Block number (from
||integer||Block instance number|
||integer||Page read latency (ms)|
||integer||Page write latency (ms)|
||integer||Log wait latency (ms)|
The read latency (
page_reads column) measures
the time from when the read request is sent to the file system
thread until the read is complete and has been reported back to
the execution thread. The write latency
page_writes) is calculated in a similar
fashion. The size of the page read to or written from a Disk
Data tablespace is always 32 KB.
Log wait latency (
log_waits column) is the
length of time a page write must wait for the undo log to be
flushed, which must be done prior to each page write.
pgman_time_track_stats table was added in