Skip to Content
Author's profile photo Gi-sung Jang

[SAP IQ] The meaning of TeamBase while load jobs are running.

We can sometimes see the below message while concurrent load jobs are running heavily.


I. 12/30 00:07:37. 0001944086 TeamBase: LoadEngine Team W 174; G 4

I. 12/30 00:07:44. 0001944101 TeamBase: LoadEngine Team W 182; G 2

I. 12/30 00:07:46. 0001944094 TeamBase: LoadEngine Team W 174; G 2

These messages are logged into iqmsg by the thread team management subsystem.

The TeamBase messages are shown because of a difference of number of threads between wanted and gained while load jobs are running.

Here are explanations of meaning of TeamBase as follows.

The Meaning of TeamBase

TeamBase: <label> W <integer>; G <integer>

– <label> says which processing algorithm is requesting a thread team.

– W <integer> tells us how many threads were requested.

– G <integer> tells us how many threads were actually allocated to the thread team by the thread team manager.


Gi-Sung Jang

Assigned Tags

      You must be Logged on to comment or reply to a post.
      Author's profile photo Johan Bornman
      Johan Bornman

      So what ratio should be considered as acceptable on a busy system?


      Johan Bornman

      Author's profile photo Mark Mumy
      Mark Mumy

      Optimally, you don't want to see these messages.  You really want your loads to get all the threads that they want.  That's not always possible for a number of reasons:

      -- -iqmt too low

      -- The max_q_threads_xxxx options are too low

      -- The governor is high enough to allow a lot of users in that are using all available threads

      I don't find it acceptable to see this message, or rather to have the W and G be too terribly different.  In a real world operational system, though, you can't always control the above factors.  In the end, is the load performing within acceptable limits to the business?  If so, then the W/G ratio is quite immaterial since loads are running within tolerances.  When loads are taking too long, clearly the pool of available threads and thread options needs to be investigated to see if there's something that can help the load get more threads.

      Also, this does change when moving to IQ 16.  The thread management stuff was changed a bit there over 15 making more threads available to queries and not always reserved for operations that may or may be running (prefetch and sweeper).