IBM Support

Retrieving Information About Objects In the IFS

How To


Summary

System administrators might want to get information about the objects in the folders and directories on IBM i. That information isn't automatically aggregated into a single location. This document explains how to put that information into a format that can be used to more easily extract useful statistics.

Steps

The first step is to collect all of the information about the IFS in tabular form.  To collect this information, the system supplies a tool called RTVDIRINF.  If you want to collect information about all the folders and directories on the system, you want to omit any remote file systems.  Most of the remote file systems are obvious but NFS mounts are not.  To find any NFS mounts on your system do the following:
qsh cmd('df -m')
The output shows all mounted file systems.  The NFS mounts are in a format, SERVER:EXPORT   mountpoint 
For example, the following is clipped from the output of the command on my lab system, the entry in bold text is an NFS mount.  This information explains that the path /mikmount on this system is currently mounted over and any results collected from that path will actually be objects on a remote file server.
  /QFileSvr.400 /QFileSvr.400               0         0   -1%        0    -1%  
  /QOPT         /QOPT                       -         -    -        -      -   
  RCH750A:/MIKSWENS  /mikmount          1677720576 657921024   61%   995898    
  61%                                                                          
  /QSYS.LIB     /QSYS.LIB          1677721584 657921144   61% 127475055    61% 
An example without any NFS mounts runs like this example that omits the libraries and remote file systems:

RTVDIRINF DIR(/) INFFILEPFX(QRYIFS) INFLIB(QGPL) OMIT('/qsys.lib' '/qntc' '/qfilesvr.400' '/qopt')
RTVDIRINF takes a while to complete, so you might want to submit that to batch.  Once complete, it creates and populates two tables in QGPL called QRYIFSO and QRYIFSD that contain information about the objects and directories.  To get useful information from the tables, you need to perform an SQL join on the two tables. 
NOTE: Much of the data in this file is in Unicode. If you are using an interface to view it that does not support converting Unicode data, it might appear incorrect.  I recommend copying the following example queries into the IBM i Access Client Solutions "Run SQL Scripts" interface.
The following queries are helpful in determining what is taking up space.
 
-- REPORT THE TOTAL SIZE OF ALL OBJECTS IN THE ROOT AND QOPENSYS FILE SYSTEMS
Select (Sum(O.QEZALCSIZE) / 1048576) As Total_Size_In_MB
   From QGPL.QRYIFSO O;

 
-- REPORT THE SIZE OF EACH DIRECTORY IN DESCENDING ORDER (LARGEST FIRST)
Select (Sum(O.QEZALCSIZE) / 1048576) As Directory_Size_In_MB,
       D.QEZDIRNAM1 As Directory_Path
   From QGPL.QRYIFSD D
        Join QGPL.QRYIFSO O
           On D.QEZDIRIDX = O.QEZDIRIDX
   Group By D.QEZDIRIDX,
            QEZDIRNAM1
   Order By 1 Desc,
            2;

 
-- REPORT THE SIZE OF ALL OBJECTS IN DESCENDING ORDER OF ALLOCATED SIZE
Select
       Case D.QEZDIRNAM1
          When '/' Then D.QEZDIRNAM1
          Else D.QEZDIRNAM1 || '/' || QEZOBJNAM
       End As "Object",
       QEZOBJTYPE As "Object Type",
       QEZALCSIZE As "Allocated Size"
   From QGPL.QRYIFSO O,
        QGPL.QRYIFSD D
   Where O.QEZDIRIDX = D.QEZDIRIDX
   Order By QEZALCSIZE Desc;

 
-- REPORT THE SIZE (IN NUMBER OF LINKS) OF EACH DIRECTORY IN DECREASING SIZE
Select QEZDIRNAM1 As "Path",
       Count(*) As "Link Count"
   From QGPL.QRYIFSO O
        Inner Join QGPL.QRYIFSD D
           On O.QEZDIRIDX = D.QEZDIRIDX
   Group By QEZDIRNAM1
   Order By Count(*) Desc;

-- FIND LINKS OLDER THAN A SPECIFIED DATE
Select D.QEZDIRNAM1 As "Parent Directory",
       QEZOBJNAM As "Object Name",
       QEZOBJTYPE As "Object Type",
       QEZCRTTIM As "Creation Datetime"
   From QGPL.QRYIFSO O,
        QGPL.QRYIFSD D
   Where O.QEZDIRIDX = D.QEZDIRIDX
         And QEZCRTTIM < '2017-03-04 00:00:00.000000';

 
-- FIND LINKS NEWER THAN A SPECIFIED DATE
Select D.QEZDIRNAM1 As "Parent Directory",
       QEZOBJNAM As "Object Name",
       QEZOBJTYPE As "Object Type",
       QEZCRTTIM As "Creation Datetime"
   From QGPL.QRYIFSO O,
        QGPL.QRYIFSD D
   Where O.QEZDIRIDX = D.QEZDIRIDX
         And QEZCRTTIM > '2018-04-04 00:00:00.000000';
-- FIND ITEMS IN A SPECIFIC PATH THAT ARE LARGER THAN A SPECIFIED SIZE
Select
       Case D.QEZDIRNAM1
          When '/' Then D.QEZDIRNAM1
          Else D.QEZDIRNAM1 || '/' || QEZOBJNAM
       End As "Object",
       QEZOBJTYPE As "Object Type",
       QEZALCSIZE As "Allocated Size (Bytes)",
       QEZDTASIZE As "Object Size (Bytes)"
   From QGPL.QRYIFSO O,
        QGPL.QRYIFSD D
   Where O.QEZDIRIDX = D.QEZDIRIDX
         And Ucase(
            Case D.QEZDIRNAM1
               When '/' Then D.QEZDIRNAM1
               Else D.QEZDIRNAM1 || '/' || QEZOBJNAM
            End) Like Ucase('/QIBM/PRODDATA/ACCESS/ACS/%') -- Modify this path accordingly
         And QEZDTASIZE > 10000;

-- FIND ALL SYMBOLIC LINKS
Select
       Case D.QEZDIRNAM1
          When '/' Then D.QEZDIRNAM1
          Else D.QEZDIRNAM1 || '/' || QEZOBJNAM
       End As "Object",
       QEZOBJTYPE As "Object Type",
       QEZALCSIZE As "Allocated Size",
       QEZDTASIZE As "Object Size (Bytes)",
       QEZCRTTIM As "Creation Datetime",
       QEZACCTIM As "Last Access Datetime",
       QEZCHGTIMA As "Last Attribute Change Datetime",
       QEZCHGTIMD As "Last Data Change Datetime",
       QEZSTGFREE As "Data Available Offline",
       QEZCHKOUT As "Object Checked Out",
       QEZCHKOWN As "User With Object Checked Out",
       QEZCHKTIM As "Checked Out Datetime",
       QEZLOCAL As "Object Stored Locally",
       QEZOWN As "Object Owner",
       QEZUID As "User ID",
       QEZOWNPGP As "Primary Group User",
       QEZGID As "Group ID Number",
       QEZAUTLST As "Authorization List",
       QEZASP As "ASP",
       QEZJRNSTS As "Journaling Status of Object",
       QEZJSUBTRE As "Journaling Subtree Semantics",
       QEZJOPTENT As "Optional Journal Entries",
       QEZJAFTERI As "Object Image After a Change is Journaled",
       QEZJBEFORI As "Object Image Before a Change is Journaled",
       QEZJRNID As "ID for Object Being Journaled",
       QEZJRNNAM As "Journal Name",
       QEZJRNLIB As "Journal Library Name",
       QEZJRNSTR As "Last Date and Time Journaling",
       QEZAUDT As "Auditing Value for Object",
       QEZBLKSIZ As "Object Block Size",
       QEZNLNK As "Number of Hard Links to Object",
       QEZFILEID As "File ID for Object",
       QEZFILEIDS As "File System Unique Object ID",
       QEZGENID As "Generation ID for File ID",
       QEZFSID As "File System ID for Object",
       QEZRDEV As "Real Device Special File",
       QEZDOM As "Domain of Object",
       QEZCRTAUD As "Auditing Value for Object in the Directory",
       QEZSCN As "Object Scan",
       QEZINHSCN As "Object Inherit Scan",
       QEZSSTATUS As "Object Scan Status",
       QEZSSIGDF As "Object Scan Signature",
       QEZSBINARY As "Scanned in Binary Mode",
       QEZSCCSID1 As "Object Scanned in the Listed CCSID1",
       QEZSCCSID2 As "Object Scanned in the Listed CCSID2",
       QEZUDATE As "Datetime Object Last Used",
       QEZUDCOUNT As "Number of Days Object Has Been Used",
       QEZURESET As "Datetime the Days Used Count was Last Reset to Zero",
       QEZPRMLNK As "Object First Name Found",
       QEZALWCKPW As "Stream File Can Be Shared with Readers and Writers",
       QEZSIG As "Object has i5/OS Digital Signature",
       QEZSYSSIG As "Signed by Trusted Source",
       QEZMLTSIG As "More Than One i5/OS Digital Signature",
       QEZDSTGOPT As "Auxiliary Storage Allocated",
       QEZMSTGOPT As "Main Storage Allocated",
       QEZDIRTYP2 As "Format of Directory",
       QEZFILTYP2 As "Format of Stream File",
       QEZUDFTYP2 As "Default File Format of Stream Files in the UDFS",
       QEZNONSAV As "Object Can Be Saved",
       QEZCLSTRSP As "Virtual disk drives for the xSeries servers",
       QEZCASE As "Case Sensitivity of the file system",
       QEZOFLOW As " Object has overflowed the ASP it resides in",
       QEZPCREAD As "Object can be changed",
       QEZPCHID As "Displayed using an ordinary directory listing",
       QEZPCSYS As "Object is a system file",
       QEZPCARC As "Object has changed since last time it was examined",
       QEZSYSARC As "Object has changed and needs to be saved",
       QEZJRCVNAM As "Oldest journal receiver to apply changes",
       QEZJRCVLIB As "Library that contains the journal receiver",
       QEZJRCVASP As "ASP name that contains the journal receiver",
       QEZJTRNI As "Current state related with commitment control",
       QEZTMPOBJ As "Object is temporary or not",
       QEZTMPUDFS As "Objects in the UDFS are temporary or not",
       QEZUNIT As "Preferred storage media for objects in UDFS"
   From QGPL.QRYIFSO O,
        QGPL.QRYIFSD D
   Where O.QEZDIRIDX = D.QEZDIRIDX
         And QEZOBJTYPE = '*SYMLNK';

-- FIND ALL OF THE JAVA CORE DUMPS. UNLESS YOU NEED THESE FOR DEBUGGING A PROBLEM, THEY ARE NOT NEEDED.
Select
       Case D.QEZDIRNAM1
          When '/' Then D.QEZDIRNAM1
          Else D.QEZDIRNAM1 || '/' || QEZOBJNAM
       End As "Object",
       QEZOBJTYPE As "Object Type",
       QEZALCSIZE As "Allocated Size"
   From QGPL.QRYIFSO O,
        QGPL.QRYIFSD D
   Where O.QEZDIRIDX = D.QEZDIRIDX
         And QEZOBJNAM Like 'core%.dmp'
Union
Select 'Total Allocated Size: ',
       '',
       Sum(QEZALCSIZE)
   From QGPL.QRYIFSO O,
        QGPL.QRYIFSD D
   Where O.QEZDIRIDX = D.QEZDIRIDX
         And QEZOBJNAM Like 'core%.dmp'
   Order By 3 Desc;
-- GENERATE SQL SCRIPT TO DELETE ALL OF THE CORE DUMPS
Select
       Case D.QEZDIRNAM1
          When '/' Then 'cl:del ''' || D.QEZDIRNAM1 || ''';'
          Else 'cl:del ''' || D.QEZDIRNAM1 || '/' || QEZOBJNAM || ''';'
       End As "Delete commands"
   From QGPL.QRYIFSO O,
        QGPL.QRYIFSD D
   Where O.QEZDIRIDX = D.QEZDIRIDX
         And QEZOBJNAM Like 'core%.dmp';

-- FIND ALL OF THE STUFF IN THE TEMPORARY DIRECTORY
Select 'Total size of junk in /tmp' As "Object",
       '' As "Object Type",
       Sum(QEZALCSIZE) As "Allocated Size (Bytes)",
       Sum(QEZDTASIZE) As "Object Size (Bytes)",
       current_date As "Creation Datetime"
   From QGPL.QRYIFSO O,
        QGPL.QRYIFSD D
   Where O.QEZDIRIDX = D.QEZDIRIDX
         And Ucase(
            Case D.QEZDIRNAM1
               When '/' Then D.QEZDIRNAM1
               Else D.QEZDIRNAM1 || '/' || QEZOBJNAM
            End) Like Ucase('/tmp/%')
Union
Select
       Case D.QEZDIRNAM1
          When '/' Then D.QEZDIRNAM1
          Else D.QEZDIRNAM1 || '/' || QEZOBJNAM
       End As "Object",
       QEZOBJTYPE As "Object Type",
       QEZALCSIZE As "Allocated Size (Bytes)",
       QEZDTASIZE As "Object Size (Bytes)",
       QEZCRTTIM As "Creation Datetime"
   From QGPL.QRYIFSO O,
        QGPL.QRYIFSD D
   Where O.QEZDIRIDX = D.QEZDIRIDX
         And Ucase(
            Case D.QEZDIRNAM1
               When '/' Then D.QEZDIRNAM1
               Else D.QEZDIRNAM1 || '/' || QEZOBJNAM
            End) Like Ucase('/tmp/%')
   Order By (3) Desc;

-- FIND ALL OF THE STUFF IN THE TEMPORARY DIRECTORY CREATED BEFORE THE LAST IPL
-- ALL OF THESE THINGS WERE CREATED AS TEMPORARY FILES BY PROCESSES NO LONGER
-- ON THE SYSTEM. THESE CAN ALL BE DELETED
Select 'Total size of junk in /tmp' As "Object",
       '' As "Object Type",
       Sum(QEZALCSIZE) As "Allocated Size (Bytes)",
       Sum(QEZDTASIZE) As "Object Size (Bytes)",
       current_date As "Creation Datetime"
   From QGPL.QRYIFSO O,
        QGPL.QRYIFSD D
   Where O.QEZDIRIDX = D.QEZDIRIDX
         And Ucase(
            Case D.QEZDIRNAM1
               When '/' Then D.QEZDIRNAM1
               Else D.QEZDIRNAM1 || '/' || QEZOBJNAM
            End) Like Ucase('/tmp/%')
         And QEZCRTTIM < (Select JOB_ENTERED_SYSTEM_TIME
               From Table (
                     QSYS2.JOB_INFO(JOB_STATUS_FILTER => '*ACTIVE', JOB_USER_FILTER => '*ALL')
                  )
               Where JOB_NAME_SHORT = 'SCPF')
Union
Select
       Case D.QEZDIRNAM1
          When '/' Then D.QEZDIRNAM1
          Else D.QEZDIRNAM1 || '/' || QEZOBJNAM
       End As "Object",
       QEZOBJTYPE As "Object Type",
       QEZALCSIZE As "Allocated Size (Bytes)",
       QEZDTASIZE As "Object Size (Bytes)",
       QEZCRTTIM As "Creation Datetime"
   From QGPL.QRYIFSO O,
        QGPL.QRYIFSD D
   Where O.QEZDIRIDX = D.QEZDIRIDX
         And Ucase(
            Case D.QEZDIRNAM1
               When '/' Then D.QEZDIRNAM1
               Else D.QEZDIRNAM1 || '/' || QEZOBJNAM
            End) Like Ucase('/tmp/%')
         And QEZCRTTIM < (Select JOB_ENTERED_SYSTEM_TIME
               From Table (
                     QSYS2.JOB_INFO(JOB_STATUS_FILTER => '*ACTIVE', JOB_USER_FILTER => '*ALL')
                  )
               Where JOB_NAME_SHORT = 'SCPF')
   Order By (3) Desc;

--Rather than always having to perform a join, you might create a view over the tables that you can then either transfer with Access Client Solutions or export to a CSV format by using CPYTOIMPF. Here is the SQL for one possible way to create that view that gives the columns meaningful names:
Create Or Replace View QGPL.QRYIFS As
      Select D.QEZDIRNAM1 As "Parent Directory",
             QEZOBJNAM As "Object Name",
             QEZOBJLEN As "Object Name Size (Bytes)",
             QEZNMCCSID As "Object Name CCSID",
             QEZREGION As "Country ID",
             QEZLANGID As "Language ID",
             QEZMODE As "File Access Mode And Type",
             QEZOBJTYPE As "Object Type",
             QEZCCSID As "Data and Extended Attributes CCSID",
             QEZALCSIZE As "Allocated Size (Bytes)",
             QEZDTASIZE As "Object Size (Bytes)",
             QEZEAs As "Number of Extended Attributes",
             QEZCEAs As "Number of Critical Extended Attributes",
             QEZEXTATRS As "Extended Attributes Size (Bytes)",
             QEZCRTTIM As "Creation Datetime",
             QEZACCTIM As "Last Access Datetime",
             QEZCHGTIMA As "Last Attribute Change Datetime",
             QEZCHGTIMD As "Last Data Change Datetime",
             QEZSTGFREE As "Data Available Offline",
             QEZCHKOUT As "Object Checked Out",
             QEZCHKOWN As "User With Object Checked Out",
             QEZCHKTIM As "Checked Out Datetime",
             QEZLOCAL As "Object Stored Locally",
             QEZOWN As "Object Owner",
             QEZUID As "User ID",
             QEZOWNPGP As "Primary Group User",
             QEZGID As "Group ID Number",
             QEZAUTLST As "Authorization List",
             QEZASP As "ASP",
             QEZJRNSTS As "Journaling Status of Object",
             QEZJSUBTRE As "Journaling Subtree Semantics",
             QEZJOPTENT As "Optional Journal Entries",
             QEZJAFTERI As "Object Image After a Change is Journaled",
             QEZJBEFORI As "Object Image Before a Change is Journaled",
             QEZJRNID As "ID for Object Being Journaled",
             QEZJRNNAM As "Journal Name",
             QEZJRNLIB As "Journal Library Name",
             QEZJRNSTR As "Last Date and Time Journaling",
             QEZAUDT As "Auditing Value for Object",
             QEZBLKSIZ As "Object Block Size",
             QEZNLNK As "Number of Hard Links to Object",
             QEZFILEID As "File ID for Object",
             QEZFILEIDS As "File System Unique Object ID",
             QEZGENID As "Generation ID for File ID",
             QEZFSID As "File System ID for Object",
             QEZRDEV As "Real Device Special File",
             QEZDOM As "Domain of Object",
             QEZCRTAUD As "Auditing Value for Object in the Directory",
             QEZSCN As "Object Scan",
             QEZINHSCN As "Object Inherit Scan",
             QEZSSTATUS As "Object Scan Status",
             QEZSSIGDF As "Object Scan Signature",
             QEZSBINARY As "Scanned in Binary Mode",
             QEZSCCSID1 As "Object Scanned in the Listed CCSID1",
             QEZSCCSID2 As "Object Scanned in the Listed CCSID2",
             QEZUDATE As "Datetime Object Last Used",
             QEZUDCOUNT As "Number of Days Object Has Been Used",
             QEZURESET As "Datetime the Days Used Count was Last Reset to Zero",
             QEZPRMLNK As "Object First Name Found",
             QEZALWCKPW As "Stream File Can Be Shared with Readers and Writers",
             QEZSIG As "Object has i5/OS Digital Signature",
             QEZSYSSIG As "Signed by Trusted Source",
             QEZMLTSIG As "More Than One i5/OS Digital Signature",
             QEZDSTGOPT As "Auxiliary Storage Allocated",
             QEZMSTGOPT As "Main Storage Allocated",
             QEZDIRTYP2 As "Format of Directory",
             QEZFILTYP2 As "Format of Stream File",
             QEZUDFTYP2 As "Default File Format of Stream Files in the UDFS",
             QEZNONSAV As "Object Can Be Saved",
             QEZCLSTRSP As "Virtual disk drives for the xSeries servers",
             QEZCASE As "Case Sensitivity of the file system",
             QEZOFLOW As " Object has overflowed the ASP it resides in",
             QEZPCREAD As "Object can be changed",
             QEZPCHID As "Displayed using an ordinary directory listing",
             QEZPCSYS As "Object is a system file",
             QEZPCARC As "Object has changed since last time it was examined",
             QEZSYSARC As "Object has changed and needs to be saved",
             QEZJRCVNAM As "Oldest journal receiver to apply changes",
             QEZJRCVLIB As "Library that contains the journal receiver",
             QEZJRCVASP As "ASP name that contains the journal receiver",
             QEZJTRNI As "Current state related with commitment control",
             QEZTMPOBJ As "Object is temporary or not",
             QEZTMPUDFS As "Objects in the UDFS are temporary or not",
             QEZUNIT As "Preferred storage media for objects in UDFS"
         From QGPL.QRYIFSO O,
              QGPL.QRYIFSD D
         Where O.QEZDIRIDX = D.QEZDIRIDX;
You could then create a CSV file on IBM i by using the CPYTOIMPF command:  CPYTOIMPF FROMFILE(QGPL/QRYIFS) TOSTMF('/QRYIFS.CSV') MBROPT(*REPLACE) RCDDLM(*CRLF) ADDCOLNAM(*SQL)
At 7.3 and later releases of IBM i, you can also use SQL Service QSYS2.IFS_OBJECT_STATISTICS to get similar information. The following SQL example counts the stream file objects by subdirectory. It includes the total size of all stream files in each subdirectory and orders the results by largest object count, descending.

 Note: This query does not run quickly.  It is meant to be run occasionally and not during a critical performance period.
with all_ifs_dirs as
(select path_name as dir_path
  from table (
      qsys2.ifs_object_statistics(
        start_path_name => '/',
        subtree_directories => 'YES',
        object_type_list => '*ALLDIR *NOQSYS',
        ignore_errors => 'YES'
        )) )
 select dir_path, count(*) as stmf_count, sum(data_size) as total_size from
  all_ifs_dirs, table ( qsys2.ifs_object_statistics(
        start_path_name => dir_path,
        subtree_directories => 'NO',
        object_type_list => '*STMF',
        ignore_errors => 'YES'
        ))
   group by dir_path     
   order by 2 desc ;
It is also suggested you put the results into a file and query the file for further analysis to not require the SQL service be ran multiple times.

[{"Type":"MASTER","Line of Business":{"code":"LOB68","label":"Power HW"},"Business Unit":{"code":"BU070","label":"IBM Infrastructure"},"Product":{"code":"SWG60","label":"IBM i"},"ARM Category":[{"code":"a8m0z0000000CGvAAM","label":"Integrated File System"}],"ARM Case Number":"","Platform":[{"code":"PF012","label":"IBM i"}],"Version":"All Versions"}]

Document Information

Modified date:
17 June 2024

UID

ibm10725583

Page Feedback