Jan Wieck wrote:
> On 5/17/2010 10:07 AM, bugzilla-daemon at main.slony.info wrote:
>> http://www.slony.info/bugzilla/show_bug.cgi?id=123

>>
>> Then 
>> -The slon for node 4 was killed
>> -Node 4 was dropped from replication, with node 1 being the event node
> 
> Was this event waited on to be confirmed by all nodes?


If I do

DROP NODE(id=4,event node=1);
sync(id=1);
wait for event(origin=1, confirmed=all, wait on=1);

still gives me the -1 error,

DROP NODE(id=4,event node=1);
wait for event(origin=1, confirmed=all, wait on=1);

'seems' to not cause the error.


> 
> 
> Jan
> 
> 
>> -A 'store node(id=4,event node=3) was executed  + add paths
>> -The slon for node 4 was started
>> -Node 4 is subscribed to the set coming from node 3.
>>
>>
>> The slon for node 4 encounters
>>
>> Received event #1 from 5000000049 type:ENABLE_SUBSCRIPTION
>> 2010-05-17 09:41:51 EDTINFO   copy_set 1 - omit=f - bool=0
>> 2010-05-17 09:41:51 EDTINFO   omit is FALSE
>> 2010-05-17 09:41:51 EDTERROR  remoteWorkerThread_1: provider node -1 for set
>> 1not found in runtime configuration
>>
>>
>> The sl_event table on node 4 contains
>>
>>
>> 4 5000000001  STORE_PATH
>> 1 5000000048  SYNC
>> 5 5000000004  SYNC
>> 3 5000000011  STORE_NODE
>> 3 5000000012  ENABLE_NODE
>> 3 5000000013  STORE_PATH
>> 3 5000000014  SUBSCRIBE_SET (3,4)
>> 3 5000000015  SYNC
>> 4 5000000002  SYNC
>> 1 5000000049  ENABLE_SUBSCRIPTION  (3,4)
>>
>> Slon seems to be processing the ENABLE_SUBSCRIPTION that is marked as having an
>> ev_origin of node 1 (the set origin) before the SUBSCRIBE_SET (having an
>> ev_origin of node 3, the provider)
>>
> 
> 


-- 
Steve Singer
Afilias Canada
Data Services Developer
416-673-1142