Previous Topic: SMP Informational ReportNext Topic: Return Code Summary Information


Miscellaneous Diagnostic Information Report

The Miscellaneous Diagnostic Information Report can be useful in dealing with CA Support and in some tuning. Similar to the Internal Statistics dynamic system table, it provides statistics for the entire MUF run. This report is available to SQL and non-SQL sites. The first section of the report provides multi-block read information. Multi-block reads can be done during sequential commands GETIT and GETPS based upon the User Requirements Table option SEQBUFS= and the contents of the actual table(s) being read. It is also done during some ADDITs when using the next overflow blocks. Multi-block read is also done (under some conditions) during Compound Boolean Selection processing. This section of the report provides the number of times various numbers of DASD blocks were read with a single EXCP. Some control exists for GETIT and GETPS, and this information is provided for information and possible tuning.

The next section of the report provides multi-block write information. With pipelining, it is possible to have multiple blocks in an area that needs to be written. When possible, CA Datacom/DB writes them together with one EXCP. This usually speeds I/O processing. These counts are provided for information and cannot be tuned.

The last section reflects miscellaneous information about the MUF execution. The sample shows one counter in this last section, BREAKS DONE. For an explanation of BREAKS DONE and other counters that could appear (if they are non-zero), see the descriptions that follow the sample.

Note: Items in the report which have a zero count are skipped.

MISCELLANEOUS DIAGNOSTIC INFORMATION EVENT COUNT WRITE BLOCKS 1 6 WRITE BLOCKS 2 11 WRITE BLOCKS 3 2 WRITE BLOCKS 4 9 WRITE BLOCKS 5 3 WRITE BLOCKS 6 10 WRITE BLOCKS 13 1 WRITE BLOCKS 14 1 WRITE BLOCKS 15 1 WRITE BLOCKS 17 4 WRITE BLOCKS 18 1 WRITE BLOCKS 19 5 WRITE BLOCKS 20 5 WRITE BLOCKS 21 2 WRITE BLOCKS 32 1 WRITE BLOCKS 36 1 WRITE BLOCKS 39 5 WRITE BLOCKS 41 1 WRITE BLOCKS 44 1 WRITE BLOCKS 46 5 WRITE BLOCKS 48 28 BREAKS DONE 1,189

This page of the report shows the following:

The following explains the possible counters that could appear in the last section of the previously shown report, if those counters are not zero. Counters with a value of zero are not shown in the report. The only counter displayed in the previously shown report is BREAKS DONE.

BREAKS DONE

If any break was done (in support of the BREAK MUF startup option), it is noted here. A break represents a time when a lower priority task which is looking at many index and or data buffers can release control and allow any waiting higher (or equal) priority task to run.

IXX QUEUE DBIDS

The maximum number of DBIDs in the index queue.

IXX QUEUE DEPTH

The maximum depth (that is, maximum number of blocks) of the index queue.

LOG VAR MAX GT8

This count is printed if you have allocated more than 8 buffers. It tells you the maximum number of buffers that have actually been used. If printed, it cannot be less than 9 nor more than the number of log-ahead buffers. It represents a high-water mark to buffer usage because the detail used counts stop at 8.

LOG VAR NO AVL

This is the count of times all log-ahead buffers are full and a wait needs to occur for a current block to free one or more buffers. If this count is large, you might want to increase the number of log-ahead buffers.

LOG VAR USED 1

This count is the number of times the first log-ahead buffer is activated. It occurs only when the single current buffer is full and no log-ahead buffers are in use. If activated, the buffer may contain only one record or a full block of records before it is written. A similar count is provided for blocks 2-7. A special count is provided for block 8, which is printed as 8+ because users can define more than 8 log-ahead buffers. The 8 count is the number of times the 8th or 9th to the end are used (similar to the 5+ buffer reuse counters).

LOG VAR WRITE 2

This count is the number of times the current log block was written with a chain of multiple blocks. These statistics are printed in the MUF EOJ report if they are not zero. The first counter provides for the times that two log current blocks are written with a single I/O, reported as LOG VAR WRITE 2. Additional counts are provided for writing three blocks, four blocks, five blocks or six blocks. In general, two blocks per track exist and so this represents 3 tracks of log data.

IO WAIT CXX BFR

Count of number of times a task waited on a CXX I/O that was already in process.

IO WAIT IXX BFR

Count of number of times a task waited on an IXX I/O that was already in process.

IO WAIT DXX BFR

Count of number of times a task waited on a DXX I/O that was already in process.

IO WAIT DATA BFR

Count of number of times a task waited on a Data I/O that was already in process.

IO WAIT LXX BFR

Count of number of times a task waited on an LXX I/O that was already in process.

LOCK LXX CUR RC

A task is waiting on access to current log buffer.

LOCK WAIT CXX B

A task is waiting on a specific Directory (CXX) block to become available.

LOCK WAIT LRU C

A task is waiting on the least recently used chain for a Directory (CXX) buffer.

MRDF IXX DIRECT

Count of times the MRDF facility could reference an IXX block directly without moving it to an IXX buffer.

OPEN EXIT WAITS

A task is waiting on open exit.

REBINDS DONE

Count of number of automatic plan rebinds that SQL has successfully completed.

WAIT_DBIO_24BIT

Count of number of times MUF did not start an I/O because its 24-bit memory was over the size limit specified with the MUF startup option X_IO_24BIT_SIZE.

X_IO_LOW_THRESHOLD

The count of times CA Datacom/DB constrained and started no I/O because the available memory was below the low threshold. This can occur once per 4/100th of a second in the MUF main task and any I/O subtask attempting to start an I/O.

Note: For information about the MUF startup option X_IO_LOW_THRESHOLD, see the CA Datacom/DB Database and System Administration Guide.

X_IO_HIGH_THRESHOLD

The count of times CA Datacom/DB constrained and started no I/O because the available memory was both below the high threshold and above the low threshold, and also because 30 or more I/Os were currently outstanding.

Note: For information about the MUF startup options X_IO_HIGH_THRESHOLD and X_IO_LOW_THRESHOLD, see the CA Datacom/DB Database and System Administration Guide.

X_IO_OS_THRESHOLD

The count of times CA Datacom/DB constrained and started no I/O because the outstanding I/O was 480 or more.