Isso soa como um bug. (Atualização: confirmada pelo apresentador).
Estou tentando abrir o OpenStack usando a versão beta do Autopilot. Eu estou usando todo o ferro (servidores físicos). Todos os servidores Dell PowerEdge, para serem específicos, estou usando:
Estou começando com o Ubuntu 14.04.2 em execução em um dos servidores r310 como meu MAAS.
Meu desafio é com o modo como a Dell reporta o disco para o comando "lshw". Por exemplo, um dos servidores r415 possui 4 unidades físicas conectadas a um controlador "SAS1068E". Eles são definidos como dois arrays RAID0 no controlador (totalizando 4 TB e 1,2 TB cada). Mas os "detalhes descobertos" informam o seguinte nos resultados detalhados do "lshw":
e a implantação CEPH falha neste nó com uma mensagem 'gancho falhou: "mon-relation-changed"'.
Os detalhes do arquivo unit-cep-0.log são:
2015-04-21 19:21:24 INFO unit.ceph/0.mon-relation-changed logger.go:40 ceph-mon-all start/running
2015-04-21 19:21:42 INFO unit.ceph/0.mon-relation-changed logger.go:40 Problem opening /dev/sda for reading! Error is 123.
Reading package lists... 87%t.ceph/0.mon-relation-changed logger.go:40
Reading package lists... 98%
Reading package lists... Done
Building dependency tree eph/0.mon-relation-changed logger.go:40
Reading state information... Doneh/0.mon-relation-changed logger.go:40
2015-04-21 19:21:45 INFO unit.ceph/0.mon-relation-changed logger.go:40 Traceback (most recent call last):
2015-04-21 19:21:45 INFO unit.ceph/0.mon-relation-changed logger.go:40 File "/usr/sbin/ceph-disk", line 2768, in <module>
2015-04-21 19:21:45 INFO unit.ceph/0.mon-relation-changed logger.go:40 main()
2015-04-21 19:21:45 INFO unit.ceph/0.mon-relation-changed logger.go:40 File "/usr/sbin/ceph-disk", line 2746, in main
2015-04-21 19:21:45 INFO unit.ceph/0.mon-relation-changed logger.go:40 args.func(args)
2015-04-21 19:21:45 INFO unit.ceph/0.mon-relation-changed logger.go:40 File "/usr/sbin/ceph-disk", line 1390, in main_prepare
2015-04-21 19:21:45 INFO unit.ceph/0.mon-relation-changed logger.go:40 zap(args.data)
2015-04-21 19:21:45 INFO unit.ceph/0.mon-relation-changed logger.go:40 File "/usr/sbin/ceph-disk", line 1013, in zap
2015-04-21 19:21:45 INFO unit.ceph/0.mon-relation-changed logger.go:40 with file(dev, 'wb') as dev_file:
2015-04-21 19:21:45 INFO unit.ceph/0.mon-relation-changed logger.go:40 IOError: [Errno 123] No medium found: '/dev/sda'
2015-04-21 19:21:45 ERROR unit.ceph/0.juju-log cmd.go:247 mon:9: Unable to initialize device: /dev/sda
2015-04-21 19:21:45 INFO unit.ceph/0.mon-relation-changed logger.go:40 Traceback (most recent call last):
2015-04-21 19:21:45 INFO unit.ceph/0.mon-relation-changed logger.go:40 File "/var/lib/juju/agents/unit-ceph-0/charm/hooks/mon-relation-changed", line 312, in <module>
2015-04-21 19:21:45 INFO unit.ceph/0.mon-relation-changed logger.go:40 hooks.execute(sys.argv)
2015-04-21 19:21:45 INFO unit.ceph/0.mon-relation-changed logger.go:40 File "/var/lib/juju/agents/unit-ceph-0/charm/hooks/charmhelpers/core/hookenv.py", line 508, in execute
2015-04-21 19:21:45 INFO unit.ceph/0.mon-relation-changed logger.go:40 self._hooks[hook_name]()
2015-04-21 19:21:45 INFO unit.ceph/0.mon-relation-changed logger.go:40 File "/var/lib/juju/agents/unit-ceph-0/charm/hooks/mon-relation-changed", line 196, in mon_relation
2015-04-21 19:21:45 INFO unit.ceph/0.mon-relation-changed logger.go:40 reformat_osd(), config('ignore-device-errors'))
2015-04-21 19:21:45 INFO unit.ceph/0.mon-relation-changed logger.go:40 File "/var/lib/juju/agents/unit-ceph-0/charm/hooks/ceph.py", line 327, in osdize
2015-04-21 19:21:45 INFO unit.ceph/0.mon-relation-changed logger.go:40 osdize_dev(dev, osd_format, osd_journal, reformat_osd, ignore_errors)
2015-04-21 19:21:45 INFO unit.ceph/0.mon-relation-changed logger.go:40 File "/var/lib/juju/agents/unit-ceph-0/charm/hooks/ceph.py", line 375, in osdize_dev
2015-04-21 19:21:45 INFO unit.ceph/0.mon-relation-changed logger.go:40 raise e
2015-04-21 19:21:45 INFO unit.ceph/0.mon-relation-changed logger.go:40 subprocess.CalledProcessError: Command '['ceph-disk-prepare', '--fs-type', u'xfs', '--zap-disk', u'/dev/sda']'
returned non-zero exit status 1
2015-04-21 19:21:45 ERROR juju.worker.uniter.operation runhook.go:86 hook "mon-relation-changed" failed: exit status 1
Parece que consegui implantar todos, exceto o Ceph, sem problemas. Estou esquecendo de algo? Existe uma maneira de dizer ao juju para ignorar ou não usar / dev / sda, / dev / sdb e / dev / sde?