Hi All,
I'm just writing up the small overlay section on slapo-syncprov (not much to it), but the man page has confused me:
syncprov-sessionlog <ops> Specify a session log for recording information about write operations made on the database. The <ops> specifies the number of operations that are recorded in the log. All write operations (except Adds) are recorded in the log. When using the session log, it is helpful to set an eq index on the entryUUID attribute in the underlying database.
"All write operations (except Adds) are recorded in the log."
What log are we talking about here?
Am Mittwoch 12 November 2008 17:27:30 schrieb Gavin Henry:
Hi All,
I'm just writing up the small overlay section on slapo-syncprov (not much to it), but the man page has confused me:
syncprov-sessionlog <ops> Specify a session log for recording information about
write operations made on the database. The <ops> specifies the number of operations that are recorded in the log. All write operations (except Adds) are recorded in the log. When using the session log, it is helpful to set an eq index on the entryUUID attribute in the underlying database.
"All write operations (except Adds) are recorded in the log."
What log are we talking about here?
The sessionlog :). IIRC it is just an in-memory list of entryUUIDs of changed (modified/deleted) entries.
What log are we talking about here?
The sessionlog :). IIRC it is just an in-memory list of entryUUIDs of changed (modified/deleted) entries.
Ah, in-memory. That reads like it's a physical log. Should we add in-memory log to avoid confusion for others like me? ;-)
On Wed, 12 Nov 2008, Gavin Henry wrote:
Ah, in-memory. That reads like it's a physical log. Should we add in-memory log to avoid confusion for others like me? ;-)
That could be clearer (I don't remember being confused by this, but lots of times what I read in the man pages is foreshadowed on the lists). Maybe "Specify a session log for recording information about write" could be "Configures an in-memory session log for..."?
----- "Aaron Richton" richton@nbcs.rutgers.edu wrote:
On Wed, 12 Nov 2008, Gavin Henry wrote:
Ah, in-memory. That reads like it's a physical log. Should we add
in-memory log to avoid
confusion for others like me? ;-)
That could be clearer (I don't remember being confused by this, but lots of times what I read in the man pages is foreshadowed on the lists). Maybe "Specify a session log for recording information about write" could be
"Configures an in-memory session log for..."?
Yes, I would prefer this.
Gvain.
Gavin Henry wrote:
Hi All,
I'm just writing up the small overlay section on slapo-syncprov (not much to it), but the man page has confused me:
syncprov-sessionlog<ops> Specify a session log for recording information about write operations made on the database. The<ops> specifies the number of operations that are recorded in the log. All write operations (except Adds) are recorded in the log. When using the session log, it is helpful to set an eq index on the entryUUID attribute in the underlying database.
"All write operations (except Adds) are recorded in the log."
What log are we talking about here?
Come on Gavin, read the docs you already wrote.
http://www.openldap.org/doc/admin24/replication.html
17.2.1.2 Syncrepl Details
... The syncrepl engine utilizes both the present phase and the delete phase of the refresh synchronization. It is possible to configure a per-scope session log in the provider server which stores the entryUUIDs of a finite number of entries deleted from a replication content. Multiple replicas of single provider content share the same per-scope session log. The syncrepl engine uses the delete phase if the session log is present and the state of the consumer server is recent enough that no session log entries are truncated after the last synchronization of the client. The syncrepl engine uses the present phase if no session log is configured for the replication content or if the consumer replica is too outdated to be covered by the session log. The current design of the session log store is memory based, so the information contained in the session log is not persistent over multiple provider invocations. ...
----- "Howard Chu" hyc@symas.com wrote:
Come on Gavin, read the docs you already wrote.
Well, as a mere user this time and not an author (I never wrote the original content, only moved it about), I'd forgotten what I'd previously read. So as we always tell our users, the Admin Guide is not the authoritative source and that they should always read the man pages first. The man page isn't clear which I will now fix.
Thanks,
Gavin.
openldap-software@openldap.org