I get occasional failures at cleanupNodeLock:

localListenThread: "select "%namespace%".cleanupNodelock(); insert
into "%namespace%".sl_nodelock values (    3, 0,
"pg_catalog".pg_backend_pid()); " - ERROR:  duplicate key violates
unique constraint "sl_nodelock-pkey"

i get to clean the stale lock by hand. from a brief glance i see no
obvious reason why the function will fail.

setup is with slony1-1.1.5 and postgresql 7.4, 1master/2 slaves. but
so far the above happened only on the one of the slaves. the setup is
as follows:

m1
master
runs slon for slavedb on m2
runs slon for master

m2
slavedb
<slon for this database is actually running on m1>

m3
slavedb
runs slon for slavedb

do you think that setup might've been causing the problems?
any hints would be appreciated!

thanks in advance,
l