Georg Thome wrote:

> thank you very much for your tips. I'm sure they would reduce 
> the time 
> for our procedure a lot.
> But there are a few other questions...
> We've got appr. 20 tables which are collecting measurement data from 
> tester maschines. If we are now running the update statistics process 
> this makes an exclusive lock on the specified table there. This 
> measurement result tables are very huge and the process takes 
> a lot of 
> time for each of them.
> I don't really understand the reason for this exclusive lock.
> Why is this exclusive lock needed, because the data and the 
> indexes are 
> not changed by only updating some system tables with 
> information about 
> the distribution and any physically address informations or whatever.
> I don't think that the information for the optimizer has to be so 
> accurate that the tables has to be locked everytime and the 
> values are 
> only a snapshot for the time the procedure started. A few 
> moments later 
> after the lock is gone the data has changed anyway a lot.
> In fact these locks are the main problem we have with the 
> whole procedure.
> For the verify procedure on the other side I can understand 
> those locks.

Hi,

after some discussion the only reason we could found for this exclusive 
locks was that they should guarantee that the result of a long running 
update statistic command could be inserted into the system table without
a collison with another transaction.

But we came to the conviction that it's better to live with an unsuccessful
update statistic run than with a lot of lock collisions.

So we will remove the exclusive locks in one of the next releases.
For detailed information about this change request have a look here:
http://www.sapdb.org/webpts?wptsdetail=yes&ErrorType=1&ErrorID=1128406


Kind regards,
Holger
SAP Labs Berlin



-- 
MaxDB Discussion Mailing List
For list archives: http://lists.mysql.com/maxdb
To unsubscribe:    http://lists.mysql.com/[EMAIL PROTECTED]

Reply via email to