def __init__(self):
rpc_pb2_grpc.WorkerServicer.__init__(self)
etcdaddr = env.getenv("ETCD")
logger.info("using ETCD %s" % etcdaddr)
clustername = env.getenv("CLUSTER_NAME")
logger.info("using CLUSTER_NAME %s" % clustername)
# init etcdlib client
try:
self.etcdclient = etcdlib.Client(etcdaddr, prefix=clustername)
except Exception:
logger.error(
"connect etcd failed, maybe etcd address not correct...")
sys.exit(1)
else:
logger.info("etcd connected")
# get master ip and report port
[success, masterip] = self.etcdclient.getkey("service/master")
if not success:
logger.error("Fail to get master ip address.")
sys.exit(1)
else:
self.master_ip = masterip
logger.info("Get master ip address: %s" % (self.master_ip))
self.master_port = env.getenv('BATCH_MASTER_PORT')
self.imgmgr = imagemgr.ImageMgr()
self.fspath = env.getenv('FS_PREFIX')
self.confpath = env.getenv('DOCKLET_CONF')
self.taskmsgs = []
self.msgslock = threading.Lock()
self.report_interval = 2
self.lock = threading.Lock()
self.mount_lock = threading.Lock()
self.cons_gateway = env.getenv('BATCH_GATEWAY')
self.cons_ips = env.getenv('BATCH_NET')
logger.info("Batch gateway ip address %s" % self.cons_gateway)
logger.info("Batch ip pools %s" % self.cons_ips)
self.cidr = 32 - int(self.cons_ips.split('/')[1])
self.ipbase = ip_to_int(self.cons_ips.split('/')[0])
self.free_ips = []
for i in range(2, (1 << self.cidr) - 1):
self.free_ips.append(i)
logger.info("Free ip addresses pool %s" % str(self.free_ips))
self.gpu_lock = threading.Lock()
self.gpu_status = {}
gpus = gputools.get_gpu_status()
for gpu in gpus:
self.gpu_status[gpu['id']] = ""
self.start_report()
logger.info('TaskController init success')
def __init__(self):
rpc_pb2_grpc.WorkerServicer.__init__(self)
etcdaddr = env.getenv("ETCD")
logger.info ("using ETCD %s" % etcdaddr )
clustername = env.getenv("CLUSTER_NAME")
logger.info ("using CLUSTER_NAME %s" % clustername )
# init etcdlib client
try:
self.etcdclient = etcdlib.Client(etcdaddr, prefix = clustername)
except Exception:
logger.error ("connect etcd failed, maybe etcd address not correct...")
sys.exit(1)
else:
logger.info("etcd connected")
# get master ip and report port
[success,masterip] = self.etcdclient.getkey("service/master")
if not success:
logger.error("Fail to get master ip address.")
sys.exit(1)
else:
self.master_ip = masterip
logger.info("Get master ip address: %s" % (self.master_ip))
self.master_port = env.getenv('BATCH_MASTER_PORT')
# get worker ip
self.worker_ip = getip(env.getenv('NETWORK_DEVICE'))
logger.info("Worker ip is :%s"%self.worker_ip)
self.imgmgr = imagemgr.ImageMgr()
self.fspath = env.getenv('FS_PREFIX')
self.confpath = env.getenv('DOCKLET_CONF')
self.rm_all_batch_containers()
self.taskmsgs = []
self.msgslock = threading.Lock()
self.report_interval = 2
self.lock = threading.Lock()
self.mount_lock = threading.Lock()
self.gpu_lock = threading.Lock()
self.gpu_status = {}
gpus = gputools.get_gpu_status()
for gpu in gpus:
self.gpu_status[gpu['id']] = ""
self.start_report()
logger.info('TaskWorker init success')
logger.info("using CLUSTER_NAME %s" % G_clustername)
# get network interface
net_dev = env.getenv("NETWORK_DEVICE")
logger.info("using NETWORK_DEVICE %s" % net_dev)
ipaddr = network.getip(net_dev)
if ipaddr == False:
logger.error("network device is not correct")
sys.exit(1)
else:
logger.info("using ipaddr %s" % ipaddr)
# init etcdlib client
try:
etcdclient = etcdlib.Client(etcdaddr, prefix=G_clustername)
except Exception:
logger.error("connect etcd failed, maybe etcd address not correct...")
sys.exit(1)
mode = 'recovery'
if len(sys.argv) > 1 and sys.argv[1] == "new":
mode = 'new'
# get public IP and set public Ip in etcd
public_IP = env.getenv("PUBLIC_IP")
etcdclient.setkey("machines/publicIP/" + ipaddr, public_IP)
# do some initialization for mode: new/recovery
if mode == 'new':
# clean and initialize the etcd table
if etcdclient.isdir(""):