Viewing HSF Information as a .csv File

If you have enabled the collection and writing of HSF records to disk, details of individual transaction runs are written to comma-separated-values files, one HSF record per line.

These files are named cashsf-a.csv and cashsf-b.csv, and are located in the system directory. One of these file is the active file at any one time, and once the active file is full, or when you click the Switch button in the HSF Data section of the ESMAC Control page, records are written to the alternate file, until the maximum file size is reached or the Switch button is clicked again.

The .csv files produced in this version of the product have the following header, and contain the following fields:

#HSFVer=05;Custom=xx;CicsFiles=xx;TSQ=xx;TDQ=xx;SUBTASK_ID=yy
Type,PID,Task,Date,Time,Tran/Job,User/DDName/CC,LU/Step,Prog/DSName,Latent/ReadCount,Resp/WriteCount,
API/RewriteCount,SQL/DeleteCount,IMS/DSDeleteCount,DSType,DSAccessType,Custom01,File01,Count01,Time01,TS01,Count01,Time01,TD01,Count1,Time1

The number of Custom, File, TS and TD fields correspond to the xx values in the header, which are set according to the ES_HSF_CFG environment variable.

When setting the variable, values must be set within a permissible range, otherwise the default value is used for that field. Also, if the field name is not explicitly set within the variable, its default value is used. The ranges and defaults are:

Field name	Range	Default
CUSTOM	0-5	0
TSQ	0-20	5
TDQ	0-20	5
SUBTASK	00-FF	00

For example: ES_HSF_CFG=CUSTOM=2;TSQ=32;SUBTASK=F1 generates 2 custom fields, , 5 TSQ fields, and 5 TDQ fields and trace id x’F1’ is used to generate SUBT records.

Field types

Type

WEB - for Web services

PID

Process ID of the SEP which ran the task.

Task

A unique task number.

Date

The date the task started to run, in the format yyyymmdd.

Time

The time the task started to run, in the format hhmmssttt (where ttt denotes thousandths of a second).

Tran/Job

The Web service name

Latent/ReadCount

The time, in milliseconds (ms), between Enterprise Server receiving the request and the task beginning to run.

Resp/WriteCount

The time, in milliseconds (ms), that the task ran for (not including latency time).

SQL/DeleteCount

The time spent, in milliseconds (ms), in SQL API (EXEC SQL statements) for this task.

CPU Time

Amount of CPU time, in milliseconds (ms), the task used¹.

Mem Usage

Amount of memory usage, in kilobytes (kB)¹.

IO Read Size

Amount of data read, in kilobytes (kB)¹.

IO Write Size

Amount of data written, in kilobytes (kB)¹.

IO Read Count

Number of read system calls¹.

IO Write Count

Number of write system calls¹.

MQ Time

Time spent, in milliseconds (ms), making MQ API calls¹.

EZASOKET Time

Time spent, in milliseconds (ms), making EZASOKET API calls¹.

DSType

This field is blank for all records.

DSAccessType

This field is blank for all records.

¹ These fields will only contain values if the region is configured to output them. This can be controlled by setting the environment variable ES_HSF_RTS_MONITOR to ON or OFF.

On Linux, support for these fields is enabled by default. To disable them, set ES_HSF_RTS_MONITOR=OFF.

On UNIX, these fields are not supported.

Configurable Field Types

The following additional fields may be repeated based on the value of the ES_HSF_CFG environment variable:

Customxx

Custom data written to the record via a call to ES_WRITE_CUSTOM_HSF from the user application.

Filexx

The name of a file that the transaction accessed.

Countxx: The number of times the transaction accessed this filexx.
Timexx: The time, in milliseconds (ms), the transaction spent accessing filexx.

TSxx

The name of a temporary storage queue (TSQ) that the transaction accessed.

Countxx: The number of times the transaction accessed TSxx.
Timexx: The time, in milliseconds (ms), the transaction spent accessing TSxx.

Sample content viewed in Microsoft Excel

Sample HSF information shown in Microsoft Excel

SUBT Type

The format for SUBT types is as follows:

"SUBT",PID,Task.Seq,Trans-id/job step,Custom data

For example:

SUBT,1234,12.1,HSF1,My data

"Custom data" is written in the same way as for custom fields, that is, by calling ES_WRITE_CUSTOM_HSF or ESCSTHSF with a 1 byte trace ID.

Note: For custom fields, multiple calls to ES_WRITE_CUSTOM_HSF/ESCSTHSF with the same ID, within the same transaction will result in data being appended to the same field up to a maximum of 255 bytes. There can be up to 5 custom fields within each record.

For sub task records, only a single ID can be used. Every call to ES_WRITE_CUSTOM_HSF/ESCSTHSF with that ID will result in a new SUBT record being generated. Custom data for these records can be up to 1024 bytes in length.

To configure this feature, the value for the ES_HSF_CFG environment variable needs to contain SUBTASK=ID where ID is the unpacked hex ID that will identify traces that will be used to generate the SUBT records. Valid values are 01 - FF. If ES_HSF_CFG does not contain a SUBTASK or if SUBTASK=00 then the feature is switched off.

For example, setting the following:

ES_HSF_CFG=SUBTASK=1A

would require a call to ES_WRITE_CUSTOM_HSF/ESCSTHSF using an ID of H"1A" in order to generate a SUBT record.

Note: If custom data contains a null character then the data will be truncated.