Julian Scarfe wrote:
>> I'm getting problems with slon on Node 3 now.  Logs below.  Why does
>> slon on
>> Node 3 shutdown?
>
> I should have added, this results in data loss on Node 3.  The data
> lost is the data replicated around the time of the slon restart(s). 
> Subsequent data flows correctly.
>
> Julian
>
Hmm.  We set up an "ACCEPT_SET" event that needs to be propagated before
nodes take account of "MOVE_SET", expressly to avoid data loss at the
time of MOVE SET.  There was a race condition where data could be lost,
and ACCEPT_SET was set up specifically to cover for this.

Would it be possible to get more specifics about what data was lost? 
e.g. - what set, and from what time (e.g. - something like "between MOVE
SET for set #3 and MOVE SET for set #4")
_______________________________________________
Slony1-general mailing list
[email protected]
http://gborg.postgresql.org/mailman/listinfo/slony1-general

Reply via email to