Bmc with this Name already exists


#1

Hi all,
I’m having a difficult time diagnosing and issue that just started affecting one of my Intel AMT nuc while trying to redeploy. I’m using Maas 2.5.1 and had deployed successfully with the “Register as MAAS KVM host” option selected and after some tests i released the machine. Upon redeploying, after the event log says “Installation complete - Node disabled netboot”, the log states, “Marking node failed - Internal error while creating KVM pod”. In looking at the regiond.conf file it states, “metadataserver.api_twisted: [info] Error while creating KVM pod: {‘name’: ['Bmc with this Name already exists.’]}”. I’ve restarted all machines, let it sit over night, restarted again, logged onto the nuc in ready state but couldn’t see anything of note. I was going to next try looking around the database to see if I could find any lingering data or … Not sure where to go next. Any suggestions would be greatly appreciated. Thanks for all your hard work.


#2

I have a bit more evidence: after turning on debugging, the regiond.log now reports:
2019-03-13 11:44:36 maasserver.models.node: [debug] Unable to format event {‘hostname’: ‘hush’, ‘config’: <function Node.mark_failed.. at 0x7f00bc299048>, ‘log_logger’: <LegacyLogger ‘maasserver.models.node’>, ‘log_level’: <LogLevel=debug>, ‘log_namespace’: ‘maasserver.models.node’, ‘log_source’: None, ‘log_format’: “Node ‘{hostname}’ failed deployment with curtin config: {config()}”, ‘log_time’: 1552502676.4321852}: get_curtin_merged_config() missing 1 required positional argument: ‘node’

and then the:

2019-03-13 11:44:36 metadataserver.api_twisted: [info] Error while creating KVM pod: {‘name’: [‘Bmc with this Name already exists.’]}