Customize the Knowledge Base › Implement Changes › Modify an Existing Rule

Modify an Existing Rule

Assume that you do not want to see all the lines of evidence produced by rule R0095 for the image VMSBUXX, because you cannot change the operation of this image. To modify a rule, you need to first disable the factory version of that rule, then copy the factory version into your MYRULES.VPR file and modify it accordingly. Make sure that you define all literals used in this rule - see the section Change a Rule Value.

The following example shows how your auxiliary rules file might look.

The bulk of the new rule came from the corresponding factory rule in PSPA$EXAMPLE:PSPA$KB.VPR. The only difference is the addition of a condition to filter out VMSBUXX image records from the rules processing. The actual rule number could not simply be copied in the auxiliary rules file, as a zero in the second character is reserved.) Most of the conclusion sections were copied from their corresponding sections in SYS$SYSTEM:PSPA$MSG.TXT. The example also shows that you can make minor modifications to the factory rules, without investing significant time and effort.

To use this rule file, see the sections Build an Auxiliary Knowledge Base and Use an Auxiliary Knowledge Base for Reporting and Archiving sections.

To change an Existing Rule

• Use the following code:

  DISABLE R0095;
  Literal TD_FILE_CACHE_HITRATIO = 70 
  TD_FILE_CACHE_MISSEDIO_RATE = 5 
  TD_XQP_CACHE = 10 
  EndLiteral
  
  Rule UR095 
  XQP_Cache_hit_ratio .lt. td_file_cache_hitratio; 
  XQP_Cache_missedio_rate .ge. td_file_cache_missedio_rate; 
  File_header_Cache_HR .lt. td_file_cache_hitratio; 
  (100 - File_header_Cache_HR) / 100 * File_header_Cache_AR 
   .ge. td_file_cache_missedio_rate; 
  Disk_header_cache_size(Highest_IO_rate_disk_x) .ge. 
   Sysgen_Acp_Hdrcache; 
  Image_name(Top_Dirio_process_x) .nes. "BACKUP"; 
  Image_name(Top_Dirio_process_x) .nes. "VMSBUXX"; 
  Occurrences = td_XQP_cache;
  Evidence = 
   Sysgen_Acp_Hdrcache 
   Disk_Header_Cache_size(Highest_IO_rate_disk_x)
   File_header_Cache_HR 
   File_header_Cache_AR 
   Volume_name(top_dskio_proc_topdsk_x)
   Disk_IO_rate(top_dskio_proc_topdsk_x)
   User_name(top_dskio_process_x)
   Image_name(top_dskio_process_x)
   Time;
  Conclusion 
  "There are too many disk I/Os caused by a low hit ratio on 
  the file header cache. This will occur if your workload 
  causes disk files to be scanned instead of repeatedly 
  accessed (i.e., BACKUP, DIR, SEARCH, etc). However, if your 
  workload does not scan disk files, so that there is still 
  useful information in the cache, then you may benefit by 
  using AUTOGEN with the feedback mechanism to automatically 
  increase the SYSGEN parameter ACP_HDRCACHE. After 
  successive uses of AUTOGEN, its feedback mechanism provides 
  the system with sufficient file header cache for the average 
  workload."
  Brief_conclusion 
  "Low hit ratio, high attempt rate on the file header cache."
  EndRule