Re: [PERFORM] Interaction between Free Space Map an alternate location

2004-11-18 Thread Dave Cramer
Ok, so the global part of the fsm is just that it is in shared memory.  If certain databases have more free space they will simply take up more of the fsm. There is no cross database movement of tuples. ( I realized this when I tried to form my next question) Dave Tom Lane wrote: Dave Cra

Re: [PERFORM] Interaction between Free Space Map an alternate location

2004-11-18 Thread Dave Cramer
No, have a look at the create database command there is a clause 'with location' that allows you to set up a separate location for the db Dave Josh Berkus wrote: Dave, Given that the free space map is global, and it is ostensibly managing free disk space, how does it deal with

Re: [PERFORM] Interaction between Free Space Map an alternate location for a database

2004-11-18 Thread Tom Lane
Dave Cramer <[EMAIL PROTECTED]> writes: > Can someone explain how the free space map deals with alternate database > locations? It doesn't really care. It identifies tables by database OID+table OID, and where they happen to sit physically doesn't matter. regards, tom la

Re: [PERFORM] Interaction between Free Space Map an alternate location for a database

2004-11-18 Thread Josh Berkus
Dave, > Given that the free space map is global, and it is ostensibly managing > free disk space, how does it deal with tuples across disk locations ? Are you talking Tablespaces? -- Josh Berkus Aglio Database Solutions San Francisco ---(end of broadcast)---

[PERFORM] Interaction between Free Space Map an alternate location for a database

2004-11-18 Thread Dave Cramer
Can someone explain how the free space map deals with alternate database locations? Given that the free space map is global, and it is ostensibly managing free disk space, how does it deal with tuples across disk locations ? Dave -- Dave Cramer http://www.postgresintl.com 519 939 0336 ICQ#14675