Tue Nov 18 12:19:51 PST 2008
- Previous message: [Slony1-general] sl_status what does it mean?
- Next message: [Slony1-general] Do finishTableAfterCopy and ANALYZE need to be serialized with data copy?
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Geoffrey <lists at serioustechnology.com> writes: > Can someone tell me what is meant by the following data from sl_status? > > st_origin | st_received | st_last_event | st_last_event_ts > | st_last_received | st_last_received_ts | > st_last_received_event_ts | st_lag_num_events | st_lag_time > -----------+-------------+---------------+----------------------------+------------------+---------------------------+----------------------------+-------------------+------------- > 1 | 2 | 73226 | 11/18/2008 09:52:03.588658 > | 73226 | 11/18/2008 09:54:45.17624 | 11/18/2008 > 09:52:03.588658 | 0 | @ 2.25 secs > (1 row) > > That is, 'st_origin' == 1. Does 1 indicate a particular status? Same > for 'st_received', what does the value 2 tell me? > > Is 2.25 secs st_lag_time good? > > Thanks for any info. As per the documentation... This view shows the local nodes last event sequence number and how far all remote nodes have processed events. - st_origin is the ID of the origin node for events, and will always be the node ID of the node where this query is being run. - st_received is the ID of the subscriber node. - st_last_event indicates what is the last event processed on the subscriber. - st_last_event_ts indicates the date that event was generated (which will be somewhat in the past) You can look in src/backend/slony1_funcs.sql to see more details about how the data is derived. If the subscriber is lagging by 0 events (as is the case in the example), then the subscriber is clearly keeping up well. st_lag_time indicates how far back events are lagging, and it is quite likely that 2.25s is pretty reasonable. There are three scenarios that seem interesting, offhand: 1. If updates are going in on the origin more or less continually, then lag time will likely be pretty low, as long as replication is keeping up. 2. If the origin sees long periods where there are no updates, lag time could get pretty high, but you may still be in a position of having no "lagging events." That's what I see in your example. 3. It could be that there's a lot of load, and that replication is lagging behind. In *that* case, you'll see the event number lag go up (e.g. - st_lag_num_events), as well as lag time (st_lag_time). That's not what I see in your example. -- select 'cbbrowne' || '@' || 'linuxfinances.info'; http://www3.sympatico.ca/cbbrowne/sgml.html "How should I know if it works? That's what beta testers are for. I only coded it." (Attributed to Linus Torvalds, somewhere in a posting)
- Previous message: [Slony1-general] sl_status what does it mean?
- Next message: [Slony1-general] Do finishTableAfterCopy and ANALYZE need to be serialized with data copy?
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
More information about the Slony1-general mailing list