Try OpenEdge Now
skip to main content
Database Administration
Reference : PROMON Utility : PROMON Block Access option
 

PROMON Block Access option

Displays database buffer pool access statistics. The following shows an example of this option's output.

Sample output for PROMON Block Access option

Block Access:
Type   Usr:Ten  Name      Domain    DB Requests       DB Reads       BI Reads      AI Reads
                                                      \Writes        \Writes       \Writes
Acc  99999      TOTAL...                  13604            72              47             0
                                                           4               3             0
Acc      0:0    dba                         682            34              47             0
                                                           3               2             0
Acc      5:1    tenant1    tenant1          607             0               0             0
                                                           0               0             0
Acc      6:3    lobtenant  lobtenant        607             0               0             0
                                                           0               0             0
Acc      7:4    lobtenant  lobtenant        647             0               0             0
                                                           0               0             0
Acc      8:2    tenant2    tenant2        647             0               0             0
                                                           0               0             0
RETURN - repeat, U - continue uninterrupted, Q - quit:
The first line displays cumulative information for all users. The six read and write columns refer to disk I/O. Reads and writes are always one block. Block size varies among systems, but is usually 512 bytes, 1,024 bytes, or 2,048 bytes.
Type
This column always has the value Acc.
Usr:Ten
The user number and tenant ID of the process.
Name
For client processes, the user name.
Domain
The user's domain name.
DB Request
The number of times the database buffer system was searched to find a block. The buffer system is searched every time a process attempts to access a record. If the block that holds the desired record is already in memory, a disk read is not required. If the ratio of DB Reqst to DB Read is not high (10 to 1 or greater), consider raising the value of the Blocks in Database Buffers (-B) startup parameter. Regardless of the number of available buffers, random record access causes a lower database request to disk read ratio than sequential record access.
DB Read
The number of database disk block reads. A database block must be read from disk when a process accesses a record whose containing block is not already in the database buffers. Recall that for read-only database requests, the OpenEdge RDBMS uses private database buffers if they are available, rather than the shared buffer pool (allocated with the Blocks in Database Buffers (-B) parameter).
DB Write
The number of database block writes to disk. Once the database buffers are full, every disk read overwrites an existing block; if the overwritten block has been modified since it was read in, it must be written to disk. This accounts for the majority of block writes.
BI Read
The number of before-image (BI) file block reads. For example, the BI file is read when a transaction is undone. The BI file has its own one-block input buffer and does not use the database buffer pool.
BI Write
The number of BI file block writes. When a record is updated, a pretransaction copy of the record is written to the BI file. When the transaction completes, the database engine writes the last BI file block out to disk (assuming you are running the database with full integrity). This post-transaction disk write accounts for the relatively high number of BI file writes, but it can be delayed with the Delay BI File Write (-Mf) startup parameter.
In addition to record images, the database engine writes to the BI file various notes and data required to reconstruct a damaged database. The BI file has its own one-block output buffer and does not use the shared database buffer pool.
AI Read
The number of after-image (AI) file block reads. The AI file is read during crash recovery. The AI file has a one-block input/output buffer and does not use the database buffer pool.
AI Write
The number of AI file block writes. When you run the database with after-imaging enabled, a copy of each note written to the BI file is written to the AI file.