Thanks Teresa.  That's a reasonable limitation for most kinds of errors I guess 
... but it would be nice to have some kind of server-wide setting to log ALL 
error sql, or per facility maybe.  Actually, sql output for the E_DM9042* 
errors (table deadlock, page deadlock, timeout ...) could even be a default 
setting, since it's pretty useful to have the sql for deadlocks.

Regards,

Jim
  ----- Original Message ----- 
  From: Teresa L. King 
  To: [email protected] 
  Sent: Wednesday, January 24, 2007 3:02 PM
  Subject: RE: [Users] Retrieving stored query text


  Hi Jim,

  Here is the answer to your second question about tracepoint sc924. Currently 
it can only be used to monitor ONE decimal error code.

  Regards,
  Teresa



------------------------------------------------------------------------------
  From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of James Gramling
  Sent: Wednesday, January 24, 2007 8:43 AM
  To: [email protected]
  Subject: [Users] Retrieving stored query text


  Hi all,

  Is there any chance that there exists a way to retrieve the stored query text 
of a named qsf query object (repeated or defined query)?Perhaps in an 
undocumented trace point or very well hidden ima mib object?

  For example: I see "Execute my_slow_query" in IPM or vdba (or as output of 
trace point SC924), and I *need* to find out what the query is doing.  I´ve 
always assumed that it was impossible to find this information without tracking 
down the user and identifying the program; but with some of the more recent 
improvements, it seems like it should be possible to query the QSF and obtain 
the original query text (or at least the qep). 

  The IMA_QSF_RQP register is tantalizingly close to exposing the information I 
want ... but it doesn´t go so far as to show the actual text.

  I´ve scanned through the source code and seen nothing obvious (and it would 
have to be very *obvious* for me to have found it! ;-) 

  Assuming such a feature does not exist, would it be possible to implement? 
(in a trace point, at least?)

  One more question: I´ve occasionally used trace point SC924 to track down 
deadlock-causing or time-out causing queries.  The behavior I have seen is that 
you can only monitor a single decimal error code at a time.  Is that correct? 
Or am I doing something wrong ? 

  TIA,

  Jim Gramling
  Interadapt
  Rio de Janeiro



------------------------------------------------------------------------------


  _______________________________________________
  Users mailing list
  [email protected]
  http://lists.ingres.com/mailman/listinfo/users
_______________________________________________
Users mailing list
[email protected]
http://lists.ingres.com/mailman/listinfo/users

Reply via email to