--On Saturday, February 09, 2008 9:27 AM -0800 James Hartley james.hartley@gmail.com wrote:
ah very good.... The documentaion I read seems to imply the number was tied to the replica, ie if the slave one used rid 001 then slave2 would use rid002.
According to your email
I would need rid001 for database 1 and rid002 for database 2 in the conf file for slave 1 and if I had another slave, say slave2 I would need to have an rid003 and rid004.
Thank you for clarifying this point.
Keep replies on the list. And no, you are incorrect still. The RIDs need to be unique per database, on a given replica. So rid001 and rid002 on replica1, rid001 and rid002 on replica2, is just fine.
--Quanah
--
Quanah Gibson-Mount Principal Software Engineer Zimbra, Inc -------------------- Zimbra :: the leader in open source messaging and collaboration
Quanah Gibson-Mount wrote:
--On Saturday, February 09, 2008 9:27 AM -0800 James Hartley james.hartley@gmail.com wrote:
ah very good.... The documentaion I read seems to imply the number was tied to the replica, ie if the slave one used rid 001 then slave2 would use rid002.
According to your email
I would need rid001 for database 1 and rid002 for database 2 in the conf file for slave 1 and if I had another slave, say slave2 I would need to have an rid003 and rid004.
Thank you for clarifying this point.
Keep replies on the list. And no, you are incorrect still. The RIDs need to be unique per database, on a given replica.
They need to be unique per *server*, not per database.
So rid001 and rid002 on
replica1, rid001 and rid002 on replica2, is just fine.
- -- Howard Chu Chief Architect, Symas Corp. http://www.symas.com Director, Highland Sun http://highlandsun.com/hyc/ Chief Architect, OpenLDAP http://www.openldap.org/project/
--On Saturday, February 09, 2008 10:45 AM -0800 Howard Chu hyc@symas.com wrote:
They need to be unique per *server*, not per database.
Oh, hm... I thought it was for helping identify a given syncrepl instance in a server with multiple DBs. Certainly in 2.3, you could use the same rid for syncrepl across multiple servers without issue. Is this a change for 2.4?
--Quanah
--
Quanah Gibson-Mount Principal Software Engineer Zimbra, Inc -------------------- Zimbra :: the leader in open source messaging and collaboration
Quanah Gibson-Mount wrote:
--On Saturday, February 09, 2008 10:45 AM -0800 Howard Chuhyc@symas.com wrote:
They need to be unique per *server*, not per database.
Oh, hm... I thought it was for helping identify a given syncrepl instance in a server with multiple DBs. Certainly in 2.3, you could use the same rid for syncrepl across multiple servers without issue. Is this a change for 2.4?
You're misunderstanding, and no, this is not a change in 2.4 relative to 2.3. "Unique per server" means every syncrepl instance in a given server must have a unique rid.
--On Saturday, February 09, 2008 11:51 AM -0800 Howard Chu hyc@symas.com wrote:
You're misunderstanding, and no, this is not a change in 2.4 relative to 2.3. "Unique per server" means every syncrepl instance in a given server must have a unique rid.
Which sounds close to what I said, excepting of course I suppose DBs with more than one syncrepl stanza in them, which would then each need their own rid. If I'm understanding you right. ;P
--Quanah
--
Quanah Gibson-Mount Principal Software Engineer Zimbra, Inc -------------------- Zimbra :: the leader in open source messaging and collaboration
Quanah Gibson-Mount wrote:
--On Saturday, February 09, 2008 11:51 AM -0800 Howard Chuhyc@symas.com wrote:
You're misunderstanding, and no, this is not a change in 2.4 relative to 2.3. "Unique per server" means every syncrepl instance in a given server must have a unique rid.
Which sounds close to what I said, excepting of course I suppose DBs with more than one syncrepl stanza in them, which would then each need their own rid. If I'm understanding you right. ;P
Correct.
openldap-software@openldap.org
-
Howard Chu -
Quanah Gibson-Mount