MVCC for massively parallel inserts - Mailing list pgsql-general

From Steven D.Arnold
Subject MVCC for massively parallel inserts
Date
Msg-id AC0C3812-3FED-11D8-B06E-000A95D9F944@neosynapse.net
Whole thread Raw
Responses Re: MVCC for massively parallel inserts  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-general
How good is Postgres' performance for massive simultaneous insertions
into the same heavily-indexed table?  Are there any studies or
benchmarks I can look at for that?

I understand Postgres uses MVCC rather than conventional locking, which
makes it easier to do parallel inserts.  In my environment, I will have
so many inserts that it is unworkable to have one machine do all the
inserting -- it would max out the CPU of even a very powerful machine,
and in any case I'd like to avoid spending that much money.  One option
is to use a cluster of commodity Intel machines running Linux or one of
the BSD's.  In many database environments, that wouldn't buy me much
because only one machine could do inserts while all the others would be
used for selects.  But I'm going to have tons of inserts and few
selects.  So I really need many of the machines in the cluster to be
able to simultaneously insert.  Is this practicable in a clustered
environment for Postgres?

Thanks in advance for any insight or references,
steve


pgsql-general by date:

Previous
From: Steven Vajdic
Date:
Subject: UNSUBSCRIBE
Next
From: Tom Lane
Date:
Subject: Re: MVCC for massively parallel inserts