Hey,

I'm trying to test failover-failback between two nodes in a typical Master-=
Slave setup.

I do not wish to completely abandon my master & then get it in sync laster =
on ( via doing a pg_dump n stuff ) in case of any issues, instead I would j=
ust do a slonik_move_set to my slave which is relatively in sync with maste=
r & start replication.

I do a slonik_move_set from node1 ( master )  to node2 ( slave ) & start re=
plication everything works fine, however i can't do a failback using slonik=
_move_set from node2 back to node1, it results into following errors which =
were completely odd IMO.

slonik_move_set --config /etc/slon_tools_failover.conf set1 1 2 | slonik

<stdin>:4: Locking down set 1 on node 1

<stdin>:5: PGRES_FATAL_ERROR select
"_replication_failover".lockSet(1); select
"_replication_failover".getMaxXid();  - ERROR:  Slony-I: set 1 is
already locked

Post manually unlocking set when tried to move set i got following error, w=
here as all the updates ran on replicated tables were received on node2

<stdin>:5: PGRES_FATAL_ERROR select "_test_cluster".moveSet(1, 2);  - ERROR=
:  Slony-I: set 1 is not subscribed by node 2

P.S. I did change the node IDs as required once a set  is moved and the Sub=
scribers are also properly updated in slony's schemas on both the nodes.

Can someone please let me know what i'm doing wrong if any.

- Manish

_________________________________________________________________
Sports, news, fashion and entertainment. Pick it all up in a package called=
 MSN India
http://in.msn.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.slony.info/pipermail/slony1-general/attachments/20090901/=
fdac27d0/attachment.htm