On Wed, 2011-01-26 at 11:07 -0600, Kevin Grittner wrote:
> > When this runs in production, how will we know whether SSI is
> > stuck
>
> Stuck? I'm not sure what you mean by that. Other than LW locking
> (which I believe is always appropriately brief, with rules for order
> of acquisition which should prevent deadlocks), the only blocking
> introduced by SSI is when there is an explicit request for
> DEFERRABLE READ ONLY transactions. Such a transaction may take a
> while to start. Is that what you're talking about?
I'm thinking of a general requirement for diagnostics.
What has been done so far looks great to me, so talking about this
subject is in no way meant to be negative. Everything has bugs and we
find them quicker if there are some ways of getting out more information
about what is happening in the guts of the system.
For HS, I put in a special debug mode and various information functions.
For HOT, I produced a set of page inspection functions.
I'm keen to have some ways where we can find out various metrics about
what is happening, so we can report back to you to check if there are
bugs. I foresee that some applications will be more likely to generate
serialization errors than others. People will ask questions and they may
claim there are bugs. I would like to be able to say "there is no bug -
look at XYZ and see that the errors you are getting are because of ABC".
> > or is consuming too much memory?
>
> Relevant shared memory is allocated at startup, with strategies for
> living within that as suggested by Heikki and summarized in a recent
> post by Jeff. It's theoretically possible to run out of certain
> objects, in which case there is an ereport, but we haven't seen
> anything like that since the mitigation and graceful degradation
> changes were implemented.
>
> > e.g. Is there a dynamic view e.g. pg_prepared_xacts,
>
> We show predicate locks in the pg_locks view with mode SIReadLock.
OK, that's good.
> > is there a log mode log_ssi_impact, etc??
>
> Don't have that. What would you expect that to show?
>
> -Kevin
-- Simon Riggs http://www.2ndQuadrant.com/books/PostgreSQL Development, 24x7 Support, Training and Services
