Re: plagued by err. 246 (even with 200K locks)
Posted in 2000
Topics: SQL Development & Query Writing, Transactions, Locking & Isolation
You could also try setting the table to row level locking if it is not set so already. Rudy Fernandes wrote: > Error 246 (Could not do an indexed read to get the next row) is not going to > go away by adding locks. It is related to lock interaction between sessions - > more than 1 session wants a resource that is already "in use" - rather than a > resource overflow. > > Secondly, isolation levels only affect locking behaviour of SELECT statements. > A row that is inserted, updated or deleted is ALWAYS locked for the duration > of the transaction. > > What you should be looking at is > > 1. Reduce, if possible, the length of each session's transactions and the > number of rows affected (and, therefore, locked). > 2. Allow sessions to wait for a locked resource (the default behaviour is to > return an error without waiting) by using the SET LOCK MODE TO WAIT statement > in the application code. (Examine the manual for details). > 3. Make sure that the tables' being hit are indexed appropriately to support > quick responses to your application requests. > > HTH > Rudy > > > > > > > > plagued by err. 246 (even with 200K locks) > > > > > > Hi, > > > We have this problem when doing a update with dirty isolation mode. > > > I have searched the archives and found out that increasing locks might > > help. > > > But it doesn't look that's helping me. Please suggest on how to overtake > > > error 246. I have also read about this being a bug. But I am not sure. > > > > > > IDS 5.10.UC3 > > > DEC OSF1 > > > > > > > > > > > > Thanks, > > > Sunil