II10841: COMMON PROBLEMS FOUND WITH IP/PRINTWAY ( 569504004 R100 )

APAR status

• Closed as canceled.

Error description

• This info apar is meant to document common problems that have
  been encountered by customers using PrintWay, and their fixes
  (where possible).
  ***************************************************************
                                INDEX
  SECTION 1 - Performance problem
  SECTION 2 - Apparent Hang, wait or loop using TCP/IP 3.2.0
  SECTION 3 - Not selecting data sets from JES spool
  SECTION 4 - JES2 performance
  SECTION 5 - INCORROUT printing to HP LaserJet printers
  SECTION 6 - LPR options file not being used
  SECTION 7 - Losing or lost Postscript data
  SECTION 8 - Number of transmission queue data sets
  SECTION 9 - Data set never prints on IBM Network Printer
  SECTION 10- ABEND02A RC0C with SYSPRINT going to sysout
  SECTION 11- MSGANFM1200I not using USERPORTS
  SECTION 12- MSGANFM1104I with errno 60 using TCP/IP 3.4.
  SECTION 13- MSGANFM1109I with errno 54 on HP printers.
  SECTION 14- ABEND09B RC0010 RC010 RC10 RSN0010 RSN010 RSN10.
  SECTION 15- ABEND047 using ACF2 security.
  SECTION 16- INCORROUT,PARTIAL OUTPUT, MISSING DATA etc...
  SECTION 17- MSGANFM131I with VALID routing file entry
  SECTION 18- MSGANFM121I ANFM121I not documeneted
  SECTION 19- MSGHASP750 $F02 RC00040006 GETCELL EXTENSION ERROR
  SECTION 20- How to get multiple copies on one printer (lpd).
  SECTION 21- Wrong font name used for -ochars parm.
  SECTION 22- msgANFM1104I with ERRNO 60 with ANY TCP/IP release
  SECTION 23- msgANFM1107I with ERRNO 1015 (eibmbadconnectionmatch
  SECTION 24- ABEND09B RC0D21 using NetSpool
  SECTION 25- Doc update for OW33273 FCB New Function
  SECTION 26- Loop in ANFFDVP with high CPU utilization
  SECTION 27- No output printed using TCP/IP 3.4 and above
  SECTION 28- ABEND001 msgIEC020I for printway message data set
  SECTION 29- ABEND0C4 IDA0192S REG9 bad with several user exits.
  SECTION 30- Printway use of Hiperspace and 2 Gig limit
  SECTION 31- Enqueue on Transmission Queue dataset
  SECTION 32- Differences between LPR/LPD and Direct Sockets
  ***************************************************************
  1 - Many customers have reported performance problems when
      printing output on one physical printer from more than 11
      PrintWay logical printers.  Often times msgANFM1103I with
      ERRNO48 (errno(48) errno 48 EADDRINUSE) is found in the
      ANFMMSG data set.  This occurs when the 11 well-known ports
      721 - 731 are exhausted and a hard coded two minute wait
      occurs (a tcp/ip restriction that is lifted in 3.2 only via
      apar PN88789).
      The fix for this is to use the USERPORTS LPR option. This
      allows ports 1024 and above to be used. In addition,
      PrintWay ptf UW37353 needs to be applied.
  2 - All users of TCP/IP 3.2.0 may see a problem that appears
      as a hang, wait, loop or incorrout with no output being
      printed. TCP/IP the ptf's for apars PQ01611, PQ06911,
      PQ03285, PQ06092 and PN88249 should ALL be applied!!
  3 - Output is available on the JES spool but is not selected
      by the JES for PrintWay.
      If using DEST, FORMS and CLASS in the PrintWay Routing
      entry the job must have matching parms and the JES2 PRTnnnn
      JES3 DEVICE(I/O) statement must have all three defined.
      Also, if using some combination of the three, make sure the
      work selection criteria (WS=) in the above JES statements
      is correct.  For example if using CLASS and FORMS only and
      NOT DEST, the work selection should be set like:
      WS=(F,Q) or some variation (as long as the R is removed).
      See the JES documentation for more info on work selection.
      Also, if using DEST=IP: in the JCL, the routing file will
      NOT be used.  IF 'R' is specified before OR after the / in
      the JES work selection criteria, a match will NEVER be
      found and the output will not be selected and 'given' to
      PrintWay.
  4 - JES2 customers: specifying TRKCELL=YES on PRTnnnn statement
      and TRKCELL=5 on SPOOLDEF is recommended for peak
      performance.
  5 - When printing to an HP LaserJet printer we have found that
      the LPR option LINETERM 0D25 must be coded to avoid
      incorrect output.  Without this the output is varied but
      most often describes as "stair-stepping".  Where each line
      is indented further than the previous line.
  6 - If using DEST=IP in the JCL for a print job, the PRTOPTNS=
      keyword must also be coded if an options file (LPR options)
      entry is to be used.
  7 - If a user is putting ascii Postscript data on the JES
      spool, there is a chance that some of the data will be
      lost.  This can occur is using BLNKTRNC=YES (jes2) or
      TRUNC=YES (jes3) is coded for the PrintWay output class.
      With these coded, jes will truncate blanks (x'40') to save
      spool space. An ascii '@' character is actually x'40' and
      is used frequently in Postscript.
      Set BLNKTRNC=NO in jes2 OUTCLASS definition or TRUNC=NO
      in jes3 SYSOUT or BUFFER definition.  YES is the default
      for both!!
  8 - There is some confusion on the number of transmission queue
      data sets that are needed in a multiple fss (multi-fss)
      environment.
      Chapter 3 of the PrintWay Guide under "ANFQUEUE DD..."
      has a NOTE that states, "Specify a different transmission
      queue data set for each PrintWay FSS."
      This should be removed (ignored).  As in the following
      section it explains that you can also share one transmis-
      sion queue data set between multiple FSS's.  How each shop
      works depends on their needs.
  9 - If printing to an IBM Network Printer (NP12 NP17 NP24
      D/T4312 D/T4317 D/T4324 4312 4317 4324) and the data set
      never prints.  Check the LPD PRINTER name that is specified
      in the LPD Port icon (\PIPE\LPDx) in the Output port window
      of the CREATE A PRINTER window of OS/2 desktop.  The LPD
      printer name MUST be TEXT for PrintWay.  This is also known
      as the print queue. (This only applies if the LPD is
      running on the OS/2 box that the printer is attached to).
      For details please see the IBM Network Printers Ethernet
      and Token Ring Configuration Guide chapter 13 and 19.
  10- ABEND02A RC0C (RC000C RC0000000C 0C 000C 0000000C
      ABEND02A-0C) in HASCHAM when a SYSPRINT DD in the
      PrintWay proc is pointing to sysout and the TCP/IP
      RESOLVER trace active.
      Please see TCP/IP apar PQ11340.  This is a tcp/ip problem
      that can be circumvented by either deactivating the
      RESOLVER trace or removing the SYSPRINT DD from the
      PrintWay proc (started task jcl).
  11- MSGANFM1200I when using a VALID print queue name and NOT
      using USERPORTS.  This error will occur when using TCP/IP
      translate tables that have NOT been created using the
      CONVXLAT program.  See the PrintWay Guide Chapter 5 under
      the heading "Using EBCDIC to ASCII Translation Tables".
      NOTE: If using a customized translate table, the PrintWay
      FSA must be drained and the FSS should be recycled before
      any changes will be seen.
  12- If using IP PrintWay with TCP/IP 3.4 and msgANFM1104I is
      issued with errno 60 (errno60 ETIMEDOUT), specify the
      USERPORTS LPR option. This assumes that the LPD supports
      ports above 1023.
  13- MSGANFM1109I with errno 54 (econnreset) can be issued
      after suffering a hardware error (output bin full) and
      using HP printers with the JetDirect card.  We found that
      they were using a PRINT QUEUE NAME in the PrintWay routing
      file entry or in the PRTQUEUE jcl keyword that is other
      than TEXT or RAW.  According to the HP "Software
      Installation Guide" for the JetDirect print server...you
      MUST use either TEXT or RAW...
      Once they changed the print queue name in PrintWay to TEXT
      they were able print with no problem.
  14- ABEND09B RC0010 may occur if running JES2 and a STEPLIB
      is present in either the JES2 proc or PRINTWAY proc (or
      both) and two copies of the jes2 load mod HASPFSSM exist
      in those libs. Using the same STEPLIB in both procs will
      resolve this error.
  15- Abend047 or Abend306 RC34 may occur using the PrintWay
      ISPF panels if using Computer Associate's ACF2 security.
      The dump will show an SVC119 (svc x'77' TESTAUTH) in acf2
      module ACF9C000. Below are the SAFDEFs needed to fix this
      for PrintWay.  The info in lower case is installation
      specific:
      INSERT SAFDEF.ipw ID(ipprtwa2) PROGRAM(ANFIEPC) RB(ANFIEPC)
      NOAPFCHK RACROUTE(REQUEST=AUTH,CLASS=DATASET,STATUS=ACCESS)
  
      INSERT SAFDEF.ipw ID(ipprtwa2) PROGRAM(ANFIEQ) RB(ANFIEQ)
      NOAPFCHK RACROUTE(REQUEST=AUTH,CLASS=DATASET,STATUS=ACCESS)
      F ACF2,REFRESH(SAFDEF) is needed after.
  16- Getting INCORROUT, partial or missing output and/or
      printer only prints the part of the output then requires
      operator intervention (printer becomes NOT READY), make
      sure that the routing file entry has a Print Queue name
      and is not specifying PORT or PORTNO.  If specifying a
      port number this implies Direct Sockets and bypasses
      LPR/LPD processing which can cause the above symptoms.
      Direct Socket support was added with a new function apar
  17- If MSGANFM131I is issued but a valid routing file DOES
      exist, it is possible that the ptf for DFSMS apar OW26115
      is not applied.  The message can be issued if a non-zero
      return code is passed back from VSAM for OPEN, GET or PUT
      macros issued to the routing data set.  The apar fixes
      some problems in this area.
      Another cause of the msgANFM131I is a mismatch between
      load libraries that are being used by the PrintWay started
      task, the TSO logon proc and what is currently running
      in MVS LLA. Make sure the SANFLOAD data set is in LINKLST
      and remove any references to it from the PrintWay proc and
      the TSO logon proc and refresh LLA. Or, assure that both
      procs point to the same copy of the data set and refresh
      LLA. Another symptom of this mismatch is msgIEC999I iec999i
  18- MSGANFM121I ANFM121I not documented.
      The IBM IP Printway Guide s544-5379-04 does not list
      the anfm121i message. The message test follows:
      ANFM121I FSS: fssname The TCP/IP INITAPI call failed, IP
      Printway is terminating
      Explanation: The INITAPI call failed. Either TCP/IP was not
      started or the name was specified incorrectly. You must
      start TCP/IP before starting IP PrintWay. In the message
      text fssname is the FSS name from FSS definition.
      System Action: The FSS terminates.
      Operator Response: No response is necessary.
      System Programmer Response: Start TCP/IP first.
      See doc apar ow36670
  19- MSGHASP750 JES2 FSI ERROR CODE= $F02 RC= 00040006 can
      occur if large RETAIN SUCCESS and/or RETAIN FAILURE times
      are used and numerous network transmission errors occur.
      Large RETAIN times mean that PrintWay cannot release the
      data sets to JES until the RETAIN times expire, which
      results in a build up of storage usage.
      Addition Symptoms: ABEND02C HASP750 RC00040006 ABENDF02
      ABEND$F02 msgANFM105I with ABEND0C4 IN ANFQSCAN.
      OW40423 will help avoid this problem too.
  20- Per doc apar OW31728 some LPD's do not support multiple
      copies via the COPIES= jcl keyword.  However, the following
      jcl may allow you to circumvent this restriction by
      generating multiple copies on the jes spool but routing it
      to the same printer:
      //PWAYJOB4 JOB ...
      //STEP1    EXEC PGM=USERA
      //OUTDS1   OUTPUT CLASS=E,FORMS=WIDE,DEST=DEPT001
      //DD1      DD SYSOUT=(,),OUTPUT=(*.OUTDS1,*.OUTDS1)
      The number of jcl refer-backs (*.OUTDS1) is directly
      related to the number of sysout data sets that get created.
  21- If using JES3 and the RECORD LPR option to pass JCL CHARS=
      parm as a -ochars= to an aix machine but the wrong font
      name is passed, add 'U' to the WS= parm of the JES3 DEVICE
      IO statement for this printer fsa.
  22- msgANFM1104I with ERRNO 60 (etimedout) running ANY TCP/IP
      release. If PRIMARYINTERFACE is not used in TCPIP.PROFILE
      data set, whatever is specified in HOME becomes the
      default host ipaddr. If HOME is other than the MVS host
      that PrintWay runs in this error will result.
      Specify PRIMARYINTERFACE if target printer is on the LAN
      that your primary interface points to OR specify
      UNRESTRICTLOWPORTS (also in TCPIP.PROFILE) and specify the
      USERPORTS LPR option in PrintWay.
      Please see tcp/ip info apars II10633 and II10899 for doc
      changes concerning the above parms.
  23- msgANFM1107I with errno1015 is caused by recycling TCP/IP
      while PrintWay remains up.  If tcpip is terminated normally
      or due to an abend, the PrintWay fss asid must be recycled
      too.  This is a permanent restriction as long as TCP/IP 3.1
      is supported.
  24- ABEND09B RC0D21 may be issued if using NetSpool with
      PrintWay.  The fix for this is tp specify "Tranmssion
      Group: Data Set" in the options file entry.  The PrintWay
      Guide states that Data Set MUST be used for Netspool output.
  25- New Function apar OW33273 adds support for using an FCB.
      However, the hierarchy for selecting an FCB name is not
      clear.  The following should be considered:
      The user can specify an FCB name in the FCB keyword in the
      JCL.  The PAGEDEF keyword can also be used. However, the FCB
      is limited to 4 characters and the PAGEDEF can be 6. If PAGE
      DEF is used only the first 4 characters will used to find
      the FCB in the library. We suggest using FCB=.  If BOTH are
      used in the JCL, PAGEDEF is given priority. If neither is in
      the JCL, the FCB named in the JES2 PRTnnnn statement or the
      NIFCB named in the PRINTDEF. In JES3 the CARRIAGE parm in
      either the DEVICE or OUTSERV statements will be used.
  26- A loop in ANFFDVP between offsets x'222' and x'760' (approx)
      and the ANFQSCAN tcb having an outstanding ENQ (SVC56, wlic=
      00020038). This is caused by the dispatching priority of GRS
      being LOWER than that of PrintWay. GRS, never gets dispatch-
      ed to work on the ENQ.
  27- If running TCP/IP 3.4 or above and not getting any output
      printed using PrintWay but the TSO LPR command works, the
      ptf's for the following apars need to be applied to tcp/ip:
      PQ29889 PQ34736 (TCP) and PQ21038 (LE). They all fix prob-
      lems with the GETHOSTNAME function (lpr cmd doesn't use).
  28- ABEND001 with msgIEC020I can occur for the printway message
      data set if it has been moved, compressed, defragged, etc.
      It should be an UNMOVABLE data set. Use DSORG=PSU in MSGINIT
      job supplied by printway to help prevent this.
  29- ABEND0C4 in IDA0192S can occur if more than one of the
      PrintWay supplied user exits are enabled and running. All
      the exits do conditional getmains of small amounts of
      storage (if getmain was done for first exit, another get-
      main will NOT be done). However, the getmain for the first
      one may not be big enough for the subsequent exits. Thus an
      overlay of VSAM control blocks can occur. To avoid this,
      make sure the first exit that gets called getmains enough
      storage area to be used by all the exits.
  30- IP Printway may use hiperspace to store a data set into
      memory before printing. This Hiperspace has a 2 gigabyte
      limit, so Printway may not be able to send datasets larger
      than 2 gig.   Hiperspace is used by LPR/LPD in all cases.
      For IPP and Direct Sockets, it is used if the maximum send
      size or delete leading and trailing formfeeds options are
      specified.
  *************************************************************
                   --- SECTION 31 ---
  When a user is in the Transmission Queue via the ISPF panels
  they receive a lock on the transmission queue dataset.
  However, this lock will be lost after 12 minutes, and at that
  point the transmission queue dataset can be accessed by others,
  which can cause the queue to become corrupted.
  *****************************************************
                 SECTION 32
  Differences between LPR/LPD protocol and Direct Sockets
  Using LPR/LPD protocol, all commands to the lpd require an
  acknowledgment. Once the data is sent, an acknowledgment from
  the LPD is required.  Using Direct Sockets protocol, there
  is no acknowledgment, and the data is handed off to TCP/IP
  to be sent to the intended port number.
  

Local fix

Problem summary

Problem conclusion

Temporary fix

Comments

• Closing
  

APAR Information

• APAR is sysrouted FROM one or more of the following:

• APAR is sysrouted TO one or more of the following:

Fix information

Applicable component levels