fb = None
+class ExceptionDoubleSSHError(Exception): pass
+
class NodeConnection:
def __init__(self, connection, node, config):
self.node = node
return "unknown"
def get_dmesg(self):
+ t_stamp = time.strftime("%Y-%m-%d-%H:%M")
self.c.modules.os.system("dmesg > /var/log/dmesg.bm.log")
- download(self.c, "/var/log/dmesg.bm.log", "log/dmesg.%s.log" % self.node)
- log = open("log/dmesg.%s.log" % self.node, 'r')
+ download(self.c, "/var/log/dmesg.bm.log", "%s/history/%s-dmesg.%s.log" % (config.MONITOR_BOOTMANAGER_LOG, t_stamp, self.node))
+ os.system("cp %s/history/%s-dmesg.%s.log %s/dmesg.%s.log" % (config.MONITOR_BOOTMANAGER_LOG, t_stamp, self.node, config.MONITOR_BOOTMANAGER_LOG, self.node))
+ log = open("%s/dmesg.%s.log" % (config.MONITOR_BOOTMANAGER_LOG, self.node), 'r')
return log
def get_bootmanager_log(self):
- download(self.c, "/tmp/bm.log", "log/bm.%s.log.gz" % self.node)
- #os.system("zcat log/bm.%s.log.gz > log/bm.%s.log" % (self.node, self.node))
- os.system("cp log/bm.%s.log.gz log/bm.%s.log" % (self.node, self.node))
- log = open("log/bm.%s.log" % self.node, 'r')
+ t_stamp = time.strftime("%Y-%m-%d-%H:%M")
+ download(self.c, "/tmp/bm.log", "%s/history/%s-bm.%s.log" % (config.MONITOR_BOOTMANAGER_LOG, t_stamp, self.node))
+ os.system("cp %s/history/%s-bm.%s.log %s/bm.%s.log" % (config.MONITOR_BOOTMANAGER_LOG, t_stamp, self.node, config.MONITOR_BOOTMANAGER_LOG, self.node))
+ log = open("%s/bm.%s.log" % (config.MONITOR_BOOTMANAGER_LOG, self.node), 'r')
return log
+
+# def get_dmesg(self):
+# self.c.modules.os.system("dmesg > /var/log/dmesg.bm.log")
+# download(self.c, "/var/log/dmesg.bm.log", "log/dmesg.%s.log" % self.node)
+# log = open("log/dmesg.%s.log" % self.node, 'r')
+# return log
+#
+# def get_bootmanager_log(self):
+# download(self.c, "/tmp/bm.log", "log/bm.%s.log.gz" % self.node)
+# #os.system("zcat log/bm.%s.log.gz > log/bm.%s.log" % (self.node, self.node))
+# os.system("cp log/bm.%s.log.gz log/bm.%s.log" % (self.node, self.node))
+# log = open("log/bm.%s.log" % self.node, 'r')
+# return log
+
def dump_plconf_file(self):
c = self.c
self.c.modules.sys.path.append("/tmp/source/")
if ret != 0:
print "\tFAILED TWICE"
#sys.exit(1)
- raise Exception("Failed twice trying to login with updated ssh host key")
+ raise ExceptionDoubleSSHError("Failed twice trying to login with updated ssh host key")
t1 = time.time()
# KILL any already running servers.
self.session = PlanetLabSession(self.hostname, False, True)
else:
self.session = PlanetLabSession(self.hostname, config.nosetup, config.verbose)
- except Exception, e:
+ except ExceptionDoubleSSHError, e:
msg = "ERROR setting up session for %s" % self.hostname
print msg
+ return False
+ except Exception, e:
traceback.print_exc()
email_exception(msg)
return False
try:
time.sleep(self.session.timeout*5)
conn = self.session.get_connection(config)
+ except EOFError:
+ # failed twice... no need to report this really, it's just in a
+ # weird state...
+ return False
except:
traceback.print_exc()
email_exception(self.hostname)
]:
sequences.update({n : "restart_bootmanager_boot"})
- # conn.restart_bootmanager('rins')
+ # conn.restart_bootmanager('reinstall')
for n in [ "bminit-cfg-auth-getplc-installinit-validate-exception-modulefail-update-debug-done",
"bminit-cfg-auth-getplc-update-installinit-validate-exception-modulefail-update-debug-done",
"bminit-cfg-auth-getplc-installinit-validate-bmexceptmount-exception-noinstall-update-debug-done",
"bminit-cfg-auth-getplc-update-installinit-validate-rebuildinitrd-netcfg-disk-update4-update3-update3-implementerror-update-debug-done",
"bminit-cfg-auth-getplc-installinit-validate-exception-bmexceptmount-exception-noinstall-update-debug-done",
"bminit-cfg-auth-getplc-update-installinit-validate-exception-bmexceptmount-exception-noinstall-update-debug-done",
+ "bminit-cfg-auth-getplc-update-installinit-validate-bmexceptvgscan-exception-noinstall-update-debug-validate-bmexceptvgscan-done",
+ "bminit-cfg-auth-getplc-update-installinit-validate-exception-noinstall-update-debug-validate-done",
]:
sequences.update({n : "restart_bootmanager_rins"})
# repair_node_keys
- sequences.update({"bminit-cfg-auth-bootcheckfail-authfail-exception-update-bootupdatefail-authfail-debug-done": "repair_node_keys"})
+ for n in ["bminit-cfg-auth-bootcheckfail-authfail-exception-update-bootupdatefail-authfail-debug-validate-exception-done",
+ "bminit-cfg-auth-bootcheckfail-authfail-exception-update-bootupdatefail-authfail-debug-done",
+ ]:
+ sequences.update({n: "repair_node_keys"})
- # conn.restart_node('rins')
+ # conn.restart_node('reinstall')
for n in ["bminit-cfg-auth-getplc-update-installinit-validate-rebuildinitrd-exception-chrootfail-update-debug-done",
"bminit-cfg-auth-getplc-update-installinit-validate-rebuildinitrd-netcfg-update3-disk-update4-exception-chrootfail-update-debug-done",
"bminit-cfg-auth-getplc-hardware-installinit-installdisk-installbootfs-installcfg-exception-chrootfail-update-debug-done",
"bminit-cfg-auth-getplc-update-installinit-validate-rebuildinitrd-netcfg-update3-implementerror-nospace-update-debug-done",
"bminit-cfg-auth-getplc-hardware-installinit-installdisk-installbootfs-exception-downloadfail-update-debug-done",
"bminit-cfg-auth-getplc-update-installinit-validate-implementerror-update-debug-done",
+ "bminit-cfg-auth-getplc-exception-update-bootupdatefail-debug-done",
]:
sequences.update({n: "restart_node_boot"})
# update_node_config_email
for n in ["bminit-cfg-exception-nocfg-update-bootupdatefail-nonode-debug-done",
"bminit-cfg-exception-update-bootupdatefail-nonode-debug-done",
+ "bminit-cfg-exception-update-bootupdatefail-nonode-debug-validate-exception-done",
"bminit-cfg-auth-bootcheckfail-nonode-exception-update-bootupdatefail-nonode-debug-done",
]:
sequences.update({n : "update_node_config_email"})
return sequence
-
def restore(sitehist, hostname, config=None, forced_action=None):
+ ret = restore_basic(sitehist, hostname, config, forced_action)
+ session.flush()
+ return ret
+
+def restore_basic(sitehist, hostname, config=None, forced_action=None):
# NOTE: Nothing works if the bootcd is REALLY old.
# So, this is the first step.
if fbnode['observed_category'] == "OLDBOOTCD":
print "\t...Notify owner to update BootImage!!!"
- if not found_within(recent_actions, 'newbootcd_notice', 3):
+ if not found_within(recent_actions, 'newbootcd_notice', 3.5):
sitehist.sendMessage('newbootcd_notice', hostname=hostname)
print "\tDisabling %s due to out-of-date BootImage" % hostname
- api.UpdateNode(hostname, {'boot_state' : 'disable'})
+ api.UpdateNode(hostname, {'boot_state' : 'disabled'})
# NOTE: nothing else is possible.
return True
debugnode = DebugInterface(hostname)
conn = debugnode.getConnection()
- #print "conn: %s" % conn
- #print "trying to use conn after returning it."
- #print conn.c.modules.sys.path
- #print conn.c.modules.os.path.exists('/tmp/source')
if type(conn) == type(False): return False
- #if forced_action == "reboot":
- # conn.restart_node('rins')
- # return True
-
boot_state = conn.get_boot_state()
if boot_state != "debug":
print "... %s in %s state: skipping..." % (hostname , boot_state)
print "...Should investigate. Skipping node."
# TODO: send message related to these errors.
- if not found_within(recent_actions, 'newbootcd_notice', 3):
+ if not found_within(recent_actions, 'baddisk_notice', 7):
+ print "baddisk_notice not found recently"
log=conn.get_dmesg().read()
sitehist.sendMessage('baddisk_notice', hostname=hostname, log=log)
- conn.set_nodestate('disable')
+ conn.set_nodestate('disabled')
return False
args['sequence'] = s
args['bmlog'] = conn.get_bootmanager_log().read()
args['viart'] = False
+ args['saveact'] = True
+ args['ccemail'] = True
sitehist.sendMessage('unknownsequence_notice', **args)
conn.restart_bootmanager('boot')
elif sequences[s] == "restart_bootmanager_rins":
print "...Restarting BootManager.py on %s "%hostname
- conn.restart_bootmanager('rins')
+ conn.restart_bootmanager('reinstall')
elif sequences[s] == "restart_node_rins":
- conn.restart_node('rins')
+ conn.restart_node('reinstall')
elif sequences[s] == "restart_node_boot":
conn.restart_node('boot')
elif sequences[s] == "repair_node_keys":
if conn.compare_and_repair_nodekeys():
# the keys either are in sync or were forced in sync.
# so try to reboot the node again.
- conn.restart_bootmanager('rins')
+ # TODO: why was this originally 'reinstall' instead of 'boot'??
+ conn.restart_bootmanager('boot')
pass
else:
# there was some failure to synchronize the keys.
args['sequence'] = s
args['bmlog'] = conn.get_bootmanager_log().read()
args['viart'] = False
+ args['saveact'] = True
+ args['ccemail'] = True
sitehist.sendMessage('unknownsequence_notice', **args)
conn.restart_bootmanager('boot')
# TODO: differentiate this and the 'nodenetwork_email' actions.
elif sequences[s] == "update_node_config_email":
- if not found_within(recent_actions, 'nodeconfig_notice', 3):
+ if not found_within(recent_actions, 'nodeconfig_notice', 3.5):
args = {}
args['hostname'] = hostname
sitehist.sendMessage('nodeconfig_notice', **args)
elif sequences[s] == "nodenetwork_email":
- if not found_within(recent_actions, 'nodeconfig_notice', 3):
+ if not found_within(recent_actions, 'nodeconfig_notice', 3.5):
args = {}
args['hostname'] = hostname
args['bmlog'] = conn.get_bootmanager_log().read()
elif sequences[s] == "update_bootcd_email":
- if not found_within(recent_actions, 'newalphacd_notice', 3):
+ if not found_within(recent_actions, 'newalphacd_notice', 3.5):
args = {}
args.update(getconf.getconf(hostname)) # NOTE: Generates boot images for the user:
args['hostname'] = hostname
# require either an exception "/minhw" or other manual intervention.
# Definitely need to send out some more EMAIL.
# TODO: email notice of broken hardware
- if not found_within(recent_actions, 'baddisk_notice', 1):
+ if not found_within(recent_actions, 'baddisk_notice', 7):
print "...NOTIFYING OWNERS OF BROKEN HARDWARE on %s!!!" % hostname
args = {}
args['hostname'] = hostname
args['log'] = conn.get_dmesg().read()
sitehist.sendMessage('baddisk_notice', **args)
- conn.set_nodestate('disable')
+ conn.set_nodestate('disabled')
elif sequences[s] == "update_hardware_email":
- if not found_within(recent_actions, 'minimalhardware_notice', 1):
+ if not found_within(recent_actions, 'minimalhardware_notice', 7):
print "...NOTIFYING OWNERS OF MINIMAL HARDWARE FAILURE on %s!!!" % hostname
args = {}
args['hostname'] = hostname
args = {}
try:
node = plccache.GetNodeByName(hostname)
- net = api.GetNodeNetworks(node['nodenetwork_ids'])[0]
+ net = api.GetInterfaces(node['interface_ids'])[0]
except:
email_exception()
print traceback.print_exc()
args['hostname'] = hostname
args['network_config'] = nodenet_str
- args['nodenetwork_id'] = net['nodenetwork_id']
+ args['interface_id'] = net['interface_id']
sitehist.sendMessage('baddns_notice', **args)