0

A report is throwing this error 

 insufficient parser memory  , during optimizer phase

I am aware of the DBSControl parameter and how it relates to this.

My Questions are

  • Best of my K, the it would be a nay... but I just wanted to check ...is there any other ODBC driver related setting that can affect this error. We know the Server DBSControl setting is there already.
  • Another hopelessly hopeful hope .....if you are not given Console privs. Is there any table in the DD out there where DBSControl settings would be stored ( like for FYI purpose ). I know it wasn't till V6 and V12. But I wondered if it got any wiser with the newer versions

So this is not getting to know the error. Pl don't explain what it means- I know what it means. My questions are specific to the above ones.

user1874594
  • 2,277
  • 1
  • 25
  • 49
  • 2
    The answer is 'no' to both questions :( – dnoeth Dec 02 '15 at 17:33
  • Thank You so much. Ok I did a small hack and managed to get DBSCONTROL info- turns out its the SAME value for Prod and Non Prod. But the very same query will explain perfectly fine in prod and an explain in non prod gets me this.. If I step back - this is happening @ the optimizer level. Optimizer conjures up the plan based on Stats. Is it poss that stats ( diff or lack of or otherwise ) on the other env cased this to happen . Any other factor can explain the disparity. Data is not the same in test and prod. – user1874594 Dec 02 '15 at 17:52
  • # AMPS are close but not the same ( 600 vs 540 ) . Besides the above ANY OTHER factors come to mind that can make it run one place and not the other. Same DB versions 14 – user1874594 Dec 02 '15 at 20:21
  • 1
    Of course different systems/stats might result in different plans, maybe it's just close to the memory limit on system #1 and a bit over the limit on #2. Are there any different (join) indexes? – dnoeth Dec 02 '15 at 20:57
  • Thx. Nope.Everything else is the same. I took prod's stats recommendations and ran them in test to see if that makes a diff ( rem. that I cant get any explain out of test ) but I am still stuck. That was chance shot. Next I'll have to compare Prod and test stats head on to see if there's some diff or getting more ( max ) stats makes it take a diff rout – user1874594 Dec 02 '15 at 21:54
  • 1
    What version of Teradata on both systems? Are the other DBS Controls the same between the two systems? Are the platforms different generations? Have you opened an incident with the GSC, it could be a known DR? Have you taken a production TSET to the pre-prod environment and tried again? – Rob Paller Dec 03 '15 at 02:41
  • Hi Rob. Yes DBS Controls values for these parameters are same. I am not sure of the platform details - but I can assume they are the same as well . DB Version and subversion is also the same ( 14.10 ). I dont have T@YS here and we have to go through System DBA's and there's good bit of tug of war goin o'ere . Even the optimizer bugs I suggested to them in past , have'nt been well heeded – user1874594 Dec 03 '15 at 20:46

0 Answers0