Lost record summary record

When the in-storage logrec data set buffer becomes filled before the recording task can be dispatched to write the stacked records to the logrec data set and remove them from the buffer, write-to-the logrec data set requests (via the RECORD service) that occur during this time are lost and cannot be written to the logrec data set. This can happen for either hardware-detected or software-detected errors. Types of errors that often result in lost records are:

Channel checks occurring continuously and so quickly that the recording task cannot keep up
Repetitive program checks in the supervisor

In both these cases, the incidents occur so close together that records cannot be written to the buffer. A count of these lost records is accumulated and later written to the logrec data set in the lost record summary (Table 1).

The lost record summary record is 25 bytes long (Table 1). The first 24 bytes is the standard software record header; byte 25 contains a count (1 to 255) of the lost records that could not be written to the logrec data set since the last lost record summary was written.

Table 1. Format of the lost record summary record
Offset		Size (bytes) alignment (bits)	Field name	Description
Dec	Hex	Size (bytes) alignment (bits)	Field name	Description
0	(0)	1	HDRTYP	Class/Source:
		.1.. 1111		Lost record summary; type=X'4F'.
1	(1)	1	HDROPRN	System/Release level:
		100. ....		OS/VS2.
		bits 3-7
		0-1F		Release level 0-31.
2	(2)	1	HDRIS	Record-independent switches:
		x... ....		Reserved.
		.1.. ....		Time-of-day (TOD) clock instruction issued. Used in conjunction with date and time values at displacement 8.
		..1. ....		Record truncated. (When EREP detects this bit being on, it does not
				edit record but prints it out in hexadecimal.)
		...1 ....		Record created by MVS/SP Version 2 or 3.
		.... 1...		TIME macro used.
		.... .xxx		Reserved.
3	(3)	3	HDRDS	Record-dependent switches:
		Byte 0
		1... ....		Short record. (Set for ‘4F’ type records to indicate that record is not
				as long as other software records.)
		.xxx xxxx		Reserved.
		Byte 1		Reserved.
		Byte 2		Reserved.
6	(6)	1	HDRCNT	Not used for lost record summary.
7	(7)	1		Reserved.
8	(8)	8	HDRTM	Time-of-day clock.
16	(10)	1	HDRCPID	Machine version code.
17	(11)	3		Central processor serial number.
20	(14)	2		Central processor machine model number.
22	(16)	2		Reserved.

				END OF STANDARD HEADER

24	(18)	1	RCBLCNT	Last field in the lost record summary. Contains the number of records that could not be written to the logrec data set.