Tech-note regarding the block-size of the QUICKEDIT recovery dataset.

Products

• Endevor
• Endevor Natural Integration
• Endevor - ECLIPSE Plugin
• Endevor - Enterprise Workbench

Issue/Introduction

If the QUICKEDIT recovery file is allocated with a System determined blocksize, a 'Recovery suspended' message, followed by a S0C3 abend will be issued.

Resolution

Two of the datasets involved in recovery in a QE/EDIT session, when RECOVERY is activated(ON), are:

1. prefix.userid.QERCVYnn.RECOVERY:
   
   This dataset contains the element as built by Endevor source management.
   
   
2. prefix.userid.CTLInnnn.BACKUP:
   
   
   • This dataset contains the changes in an internal EDIT format, as determined by the ISPF/EDITOR, as invoked by QUICKEDIT.
   • The third qualifier in the '.BACKUP' dataset depends on QE being invoked with NEWAPPL(CTLI), where CTLI is the recommended (it is required for commands like ESORT to function) application name.
     This dataset is the actual recovery dataset in ISPF/EDIT terms.
   • According to the IBM ISPF Planning and Customizing manual, this dataset has a record format U, record length 0, the default Block size is 13680, and a System determined block size is not supported.
   • A pitfall is therefore that the storage administrator implements rules that force the blocksize to zero, thus using a System determined block size, and the prefix.userid.CTLInnnn.BACKUP recovery dataset has not been excluded.
   • The result can be that a "Recovery suspended" message is displayed, and a S0C3 abend is issued by QE on SAVE.
   • To avoid such problems, it is recommended to exclude these data sets from your allocation rules coded in for example SMS and CA Allocate.