14 Replies Latest reply on Jan 29, 2013 1:33 PM by Shaz3kgt

    Forcing Extract to check point/extract looking for 5-day old archive_log

    Shaz3kgt
      Hey guys, I'm having some what of an odd issue and can not figure out what the issue is. I believe theoretically I have tracked it down but cant seem to proof it.

      Here is what happens, I have an extract process running for a specific schema, now that schema only gets inserted into every 4-5 days. CDC is able to transfer the data and is running as expected.

      Here is the issue, when I stop cdc , (patching or maintenance) and restart it lets say 20-30 minutes later it errors out.
      The error message indicates its looking for an archive log, the odd part is its looking for an archive log that is 4-5 days old. We only keep 1-day on disk and about 2-3 days in RMAN. (and up to 6 months on tape)

      I can not figure out why the extract process would want an archive log 5 days old?? The only explanation I have is since this schema only gets written too (inserted specifically) every 4-5 days that the last time the extract process actually got data from an archive log(or redo-log) was about 4-5 days ago and somewhere in the checkpoint table is that info. Now , if that is the case and I'm on the right track , what can I do?? Do I need to somehow have the extract process "checkpoint"(not sure how else to describe it) everyday?
      If my theory is wrong, any ideas of what might be causing this??