On 12/05/11 1:34 PM, C. Mundi wrote:
> So that's my concern. I'm doing 80% reads which are all non-blocking
> with 20% writes mixed in, and I need to avoid the effect of writes
> blocking queries which do not need to traverse branches affected by
> the write.
postgres does no blocking on inserts/updates. the commonest lock is if
you're doing a transaction, and need to select something prior to
updating it, then you use a SELECT ... FOR UPDATE; this locks just the
rows you're going to update so noone else can update them (but other
clients can still read the existing value prior to your COMMIT).
--
john r pierce N 37, W 122
santa cruz ca mid-left coast