On 11/12/2004 6:36 AM, Hannu Krosing wrote:

> Hi
> 
> I have a situation, where my replica on node3 is ~1 day behind reality
> (due to similar situation to the one discussed in the Slow-ny thread), but
> 
> accountdb=# select * from _accountdb_cluster.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 |        130148 | 2004-11-12 12:47:09.334286 |   
>        130147 | 2004-11-12 11:46:07.979793 | 2004-11-12
> 12:46:07.30647  |                 1 | 00:48:32.935116
>          1 |           3 |        130148 | 2004-11-12 12:47:09.334286 |   
>        130146 | 2004-11-12 11:45:05.514404 | 2004-11-12
> 12:45:04.884938 |                 2 | 00:49:35.356648
> (2 rows)
> 
> 
> I was under the impression that sl_status should somehow show how far
> behind the replica is.

a) Please send this sort of paste as attachment, the linebreaks make it
    hard to read.

b) Could it be that your origins clock is about 50 minutes off the
    subscribers?

c) The last generated event on node 1 is 130148. Node 2 has confirmed
    130147 and node 3 has confirmed 130146. Since these events don't seem
    to be that old, what gives you the impression the node would be 1 day
    behind? I see 48 and 49 minutes, which would be very odd assuming
    that all slon daemons are currently running.


Jan

> 
> ---------------
> Hannu
> 
> _______________________________________________
> Slony1-general mailing list
> Slony1-general at gborg.postgresql.org
> http://gborg.postgresql.org/mailman/listinfo/slony1-general


-- 
#======================================================================#
# It's easier to get forgiveness for being wrong than for being right. #
# Let's break this rule - forgive me.                                  #
#================================================== JanWieck at Yahoo.com #