vbr replication during standby unavailable

avkirilishin

We use vbr for replicate two similar clusters. There is one standby node per cluster.
When standby node in destination cluster take the place of a failed node (host crashes), replication fails because of rsync daemons checks and others (crashed node becomes standby and unavailable).
This is vbr's fix:

8076,8083d8075
<     @property
<     def participating_hosts(self):
<         node_subset = set(self._config._topology[src_node_name].backup_host for src_node_name in self._selected_nodes)
<         src_nodeset = set([
<             self._database_info.get_host_name(node)
<             for node in self._participating_nodes])
<         return src_nodeset | (set(self._dest_hosts.values()) & node_subset)
< 
9044c9036
<     def __init__(self, config, node_subset=None):
---


def __init__(self, config):

9047d9038 < self.node_subset = node_subset 9130,9133c9121 < if self.node_subset: < all_backup_host = set(self._config._topology[src_node_name].backup_host for src_node_name in self.node_subset) < else: < all_backup_host = set(loc.backup_host for loc in self._config.all_backup_lock_locations()) ---

    all_backup_host = set(loc.backup_host for loc in self._config.all_backup_lock_locations())

10011c9999 < RsyncDaemonLauncher(config, node_subset=opt_nodes).launch_rsync_daemons() ---

        RsyncDaemonLauncher(config).launch_rsync_daemons()

Does another way exist to replicate objects with down standby host?

SruthiA

Can you share us the exact error message thrown by vbr replicate?

SruthiA

Along with that please share vbr.log if you can replicate the issue even now.

Jim_Knicely

@avkirilishin - And can you also let us know the version of Vertica?