[elephant-devel] Re: Updated version of last Postmodern patch bundle

[Ian Eslick]

The with-transaction macro passes keyword arguments to the controller- 
create-transaction backend fn which you may also call directly if you  
want to do something more sophisticated than wrap body code in with- 
transaction.

Some examples for BDB:

dirty-read - Read written but uncommitted values.  For example if  
you're walking over a btree and you don't want to restart because  
someone inserted a value on a previously read or just about to be read  
value.  Also saves computation time and txn memory.

txn-nosync - Do not wait while for synchronization to disk.  Can have  
a noticable performance benefit but it means that the transaction is  
not durable - future code may assume it committed, but if the database  
crashes that transaction may be unwound.

txn-nowait - do not block on locks (return error DB_LOCK_NOTGRANTED  
immediately)

Ian

On Mar 16, 2008, at 11:48 AM, Alex Mizrahi wrote:

> ??>>>>>  * handle pg error 23505 (duplicate primary key race  
> condition)
> ??>>
> ??>> i guess race condition in plpgsql should be handled by  
> switching to
> ??>> serializable transaction isolation level. (that's what i'm doing)
>
> LPP> I don't quite see how the two of them go together.
> LPP> The isolation guarantee level should be settable by the user
> LPP> and the PM backend should work in any case, shouldn't it?
>
> why would we need anything other than best isolation level?
> people might prefer "read commited" because it yields no retries,  
> but we
> need retries anyway (in case of deadlock or duplicate pk), so i see no
> reason.
>
> i'm pretty sure there can be other race conditions and best way to  
> avoid
> them is to use serializable isolation level.
> also Ian says that serializable is default level in BDB backend, i  
> think it
> would be good if two backends will be consistent on this matter.
>
>
>
> _______________________________________________
> elephant-devel site list
> elephant-devel at common-lisp.net
> http://common-lisp.net/mailman/listinfo/elephant-devel