[erlang-questions] mnesia:dirty_update_counter and replicated tables

Ulf Wiger (TN/EAB) <>
Thu Sep 21 11:25:49 CEST 2006


Hi Scott,

I believe the quick'ish answer is "no".

Dirty_update_counter() is not more safe than dirty_write(),
and dirty_write() and dirty_write() is known to lead to 
inconsistencies in certain failure situations. This 
applies equally to dirty_update_counter()

BR,
Ulf W

> -----Original Message-----
> From:  
> [mailto:] On Behalf Of 
> Scott Lystig Fritchie
> Sent: den 21 september 2006 09:10
> To: 
> Subject: [erlang-questions] mnesia:dirty_update_counter and 
> replicated tables
> 
> There hasn't been any discussion of 
> mnesia:dirty_update_counter here in quite a while.
> 
> Looking through my archive, I didn't find a definite answer.  
> Is mnesia:dirty_update_counter safe to use when the underlying 'set'
> table is replicated?  If I use mnesia:dirty_update_counter on 
> the same counter on two different Mnesia nodes, would I ever 
> be unpleasantly surprised?
> 
> The flow seems to go through the transaction manager, which 
> bodes well for multi-node safety?  Umrfl, it's too late to be 
> reading Mnesia code.  It will be a good learning exercise to 
> continue my reading, but in the meantime ... is there a 
> quick'ish answer from a kind soul?
> 
> Or will a tormenting soul tell me to use tracing and the 
> 'dbg' module to trace the path for myself?  :-)
> 
> -Scott
> _______________________________________________
> erlang-questions mailing list
> 
> http://www.erlang.org/mailman/listinfo/erlang-questions
> 




More information about the erlang-questions mailing list