LOG datasets in SOLVE / Netmaster getting SD37-04 abend
Article ID: 186194
Updated On:
Products
Issue/Introduction
Upgraded to SOLVE:Operations 11.9 recently.
The RUNSYSIN is modified to write logs out to data sets instead of JES.
It shows
DD=LOG1,DISP=SHR,DSN=hlq.LOG1
DD=LOG2,DISP=SHR,DSN=hlq.LOG2
* DD=LOG3,DISP=SHR,DSN=hlq.LOG3
* DD=LOG4,SYSOUT=*,FREE=CLOSE
* DD=LOG5,SYSOUT=*,FREE=CLOSE
* DD=LOG6,SYSOUT=*,FREE=CLOSE
* DD=LOG7,SYSOUT=*,FREE=CLOSE
* DD=LOG8,SYSOUT=*,FREE=CLOSE
* DD=LOG9,SYSOUT=*,FREE=CLOSE
However, the JES log shows the log abending with
13.18.28 S0000311 IEC031I D37-04,IFG0554P,NMCS71,$$$$$$@,LOG1,20B6,NMD037, 732
732 <hlq>.LOG1
At this point, if I try to look at LOG2 (or may be enmpt on a new reigon, it has old data, but LOG1 is not logging further information.
Environment
All SOLVE and Netmaster products
Cause
Resolution
The product is working as designed.
Hardcopy logs (dataset instead of SYSOUT) are purposely defined with no secondary extents.
An SD37-04 occurs when one log fills up and the logswap process switches over to use the next log.
Here is what can be expected each time a logswap occurs for these files:
1. LOG1 has enough records written to it that it hits the end of the dataset, which causes an SD37-04 and swaps the log.
The JESMSGLG shows:
13.18.28 S0023516 IEC031I D37-04,IFG0554P,NMCS71,$$$$$$@,LOG1,20B6,NMD037, 732
732 <hlq>.LOG1
The activity log (/LOG from any command line in SOLVE) shows
13.18.28 N16103 I/O ERROR ON ACTIVITY LOG DD=LOG1 , ABEND CODE=D37-04, USING NEXT LOG.
13.18.28 N16110 LOGSWAP COMPLETED, NEW LOG DD=LOG2
At this point, if I try to look at LOG2, it may appear to either be empty or contain old information, but LOG1 is not logging further information.
So it looks like there's a problem.
However,when I go to CMD and issue SHOW LOGFILES I see
sh logfiles
N11924 -DD- STATUS- ---START TIME-- ---STOP TIME--- PAGES -SWAPPED BY-
N11923 LOG1 SWAPPED 20.069 13.16.52 20.069 13.18.28 181 FULL
N11923 LOG2 IN USE 20.069 13.18.28 37
N11907 *END*
** END OF DELIVERED MESSAGES **
So LOG2 is actually in use. The reason it initially appears to contain old/missing information is that the buffer that writes to the LOGx is not yet full.
This is generally only an issue on a test system with little traffic.
Once the buffer writes out to LOG2, the data will be visible.
This same processes and messages apply when going from LOG2 to LOG3, or from LOG3 back to LOG1.
Once all logs contain data, a logswap causes new data to overwrite the old, so you might see something like
005064 14.04.56.57 *001185* <userid> N22202 MAIVAUTH YES DEFAULT -
005065 14.04.56.57 *001185* <userid> N22202 MAXRUSZ 4096 DEFAULT -
005066 14.04.56.57 *001185* <userid> N22202 MDFYLIM 5 DEFAULT -
The SD37-04 is simply the log spin doing its thing.
Additional Information
If you need to see logging information that is not yet appearing in the hardcopy logs, you can view the log directly from the SOLVE / Netmaster region by logging on and issuing either /LOG or $LOG from any command line. This data is stored in the hlq.LOG01-LOG03 VSAM files defined in the LOGFILES parameter group.
Feedback
