2 Replies Latest reply: Nov 19, 2012 8:00 AM by Felix_GG RSS

    logdump utility: dump

    Felix_GG
      Hi guys,

      i have a question about the logdump.
      I opend a file and displayed the first record, worked fine:


      Logdump 279 >pos 0
      Reading forward from RBA 0
      Logdump 280 >n

      2012/11/15 14:35:37.020.665 Update Len 30 RBA 0
      Name: \M7.$DATA04.DVPOOMQR.MQRGOMHB
      After Image: RBA 0 (x00000000) Partition 0 G s
      0000 0000 3230 3132 2d31 312d 3135 3a31 343a 3335 | ....2012-11-15:14:35
      3a33 372e 3031 3939 3937 | :37.019997


      But if i dump this record, i see something like a header (?) before, not the expected date as the command before:

      Logdump 281 >dump 30
      Dump 30 Bytes at RBA 0
      4700 00aa 4800 0040 4500 0041 001e 0acf 02f1 f4bc | G...H..@E..A........
      2ea9 9cf9 0000 0000 06be | ..........

      If i skip the first 75 bytes, i see the expected result:

      Logdump 282 >pos 75
      Reading forward from RBA 75
      Logdump 283 >dump 30
      Dump 30 Bytes at RBA 75
      1e00 0000 0032 3031 322d 3131 2d31 353a 3134 3a33 | .....2012-11-15:14:3
      353a 3337 2e30 3139 3939 | 5:37.01999


      My question, can i supress dumping this header (?) ?
      If not, HOW can i determine the offset to use, 75 in this example ?
      I found it by analyzing the displayed date, but there is maybe a better way ?
        • 1. Re: logdump utility: dump
          MikeN
          Felix_GG wrote:
          Hi guys,

          i have a question about the logdump.
          I opend a file and displayed the first record, worked fine:


          Logdump 279 >pos 0
          Reading forward from RBA 0
          Logdump 280 >n

          2012/11/15 14:35:37.020.665 Update Len 30 RBA 0
          Name: \M7.$DATA04.DVPOOMQR.MQRGOMHB
          After Image: RBA 0 (x00000000) Partition 0 G s
          0000 0000 3230 3132 2d31 312d 3135 3a31 343a 3335 | ....2012-11-15:14:35
          3a33 372e 3031 3939 3937 | :37.019997


          But if i dump this record, i see something like a header (?) before, not the expected date as the command before:

          Logdump 281 >dump 30
          Dump 30 Bytes at RBA 0
          4700 00aa 4800 0040 4500 0041 001e 0acf 02f1 f4bc | G...H..@E..A........
          2ea9 9cf9 0000 0000 06be | ..........

          If i skip the first 75 bytes, i see the expected result:

          Logdump 282 >pos 75
          Reading forward from RBA 75
          Logdump 283 >dump 30
          Dump 30 Bytes at RBA 75
          1e00 0000 0032 3031 322d 3131 2d31 353a 3134 3a33 | .....2012-11-15:14:3
          353a 3337 2e30 3139 3939 | 5:37.01999


          My question, can i supress dumping this header (?) ?
          If not, HOW can i determine the offset to use, 75 in this example ?
          I found it by analyzing the displayed date, but there is maybe a better way ?
          Displaying records in logdump is only really possible when you are at the beginning of a record (at the start of a header), not at arbitrary points in the trail. If you position to some arbitrary location, then run "sfh" (scanforheader) and it will reposition you to the beginning of the next record, and you can look at it using "n" (next).

          (In logdump, run "help" and it will give you a summary of commands to run; the doc's also contain info.)

          But more to the question: a header will always be displayed, but it shouldn't be in the hex dump portion of the output; when logdump is properly positioned at the beginning of the record, the output will be nicely formatted as readable text. The verbosity of header info is controlled by setting things (in logdump) like "detail on", "ghdr on", "usertoken detail", "ggstoken detail", "fileheader detail" (etc). When exiting logdump, everything is reset to defaults when restarting logdump.

          Edit: re-reading the question again, rather than using "dump" (which would require knowing the length of the header in bytes), won't "next" be sufficient for printing out the record? If the entire record is not being printed, try "reclen <nnn>" to print out more data for a specific column, or "detail data" to print out more info about each column.

          Hope it helps..
          -m

          Edited by: MikeN on Nov 16, 2012 11:22 PM
          • 2. Re: logdump utility: dump
            Felix_GG
            Hi Mike,

            no, 'next' is not enough for my needs ...

            I used this 'small record' only as an example, the 'true' record i want to examine is much bigger and cointans a lot of fields.
            So, i use dump to see the value for ONE field in this record.
            For this reason i get the needed information from the definition file (size, offset) and dump a single field of the whole record.

            The positioning should be OK, anyway if i use 'next' to navigate to a record or 'pos' with the RBA value from a report-file, i have to skip this 75 Bytes first ...