From be11398f97d51c72f275df3d4f7e095e900f2add Mon Sep 17 00:00:00 2001
From: Stephen Soltesz <soltesz@cs.princeton.edu>
Date: Wed, 11 Mar 2009 21:06:44 +0000
Subject: [PATCH] split reboot.py across pcucontrol and monitor modules.
now pcucontrol is compeletely independent. improved the timeout function in
util/command.py. allows script to hop past nodes that are really
broken. minor edits to address reboot.py changes.
---
findbadpcu.py | 1 -
grouprins.py | 2 +-
monitor/reboot.py | 433 +---------------------------------------
monitor/util/command.py | 102 ++++++----
nodeinfo.py | 2 +-
nodequery.py | 3 -
pcubad.py | 2 +-
pcucontrol/reboot.py | 108 +---------
sitebad.py | 1 -
siteinfo.py | 1 -
10 files changed, 70 insertions(+), 585 deletions(-)
diff --git a/findbadpcu.py b/findbadpcu.py
index 7e84513..d00d7f7 100755
--- a/findbadpcu.py
+++ b/findbadpcu.py
@@ -13,7 +13,6 @@ import threadpool
import threading
import monitor
-from pcucontrol import reboot
from monitor import config
from monitor.database.info.model import FindbadPCURecordSync, FindbadPCURecord, session
from monitor import database
diff --git a/grouprins.py b/grouprins.py
index 433ecd3..eb89e6a 100755
--- a/grouprins.py
+++ b/grouprins.py
@@ -17,7 +17,7 @@ from monitor import util
from monitor import const
from monitor import database
from monitor import parser as parsermodule
-from pcucontrol import reboot
+from monitor import reboot
from monitor.wrapper import plc
api = plc.getAuthAPI()
diff --git a/monitor/reboot.py b/monitor/reboot.py
index c3f6103..3a8e8c5 100755
--- a/monitor/reboot.py
+++ b/monitor/reboot.py
@@ -17,8 +17,6 @@ import pcucontrol.transports.ssh.pxssh as pxssh
import pcucontrol.transports.ssh.pexpect as pexpect
import socket
-
-
# Use our versions of telnetlib and pyssh
sys.path.insert(0, os.path.dirname(sys.argv[0]))
import pcucontrol.transports.telnetlib as telnetlib
@@ -29,6 +27,8 @@ from monitor import config
from monitor.util import command
from monitor.wrapper import plc
+from pcucontrol.reboot import pcu_name, model_to_object, reboot_api, convert_oldmodelname_to_newmodelname, reboot_test_new
+
# Event class ID from pcu events
#NODE_POWER_CONTROL = 3
@@ -41,312 +41,6 @@ logger = logging.getLogger("monitor")
verbose = 1
#dryrun = 0;
-class ExceptionNoTransport(Exception): pass
-class ExceptionNotFound(Exception): pass
-class ExceptionPassword(Exception): pass
-class ExceptionTimeout(Exception): pass
-class ExceptionPrompt(Exception): pass
-class ExceptionSequence(Exception): pass
-class ExceptionReset(Exception): pass
-class ExceptionPort(Exception): pass
-class ExceptionUsername(Exception): pass
-
-
-
-# PCU has model, host, preferred-port, user, passwd,
-
-# This is an object derived directly form the PLCAPI DB fields
-class PCU(object):
- def __init__(self, plc_pcu_dict):
- for field in ['username', 'password', 'site_id',
- 'hostname', 'ip',
- 'pcu_id', 'model',
- 'node_ids', 'ports', ]:
- if field in plc_pcu_dict:
- self.__setattr__(field, plc_pcu_dict[field])
- else:
- raise Exception("No such field %s in PCU object" % field)
-
-# These are the convenience functions build around the PCU object.
-class PCUModel(PCU):
- def __init__(self, plc_pcu_dict):
- PCU.__init__(self, plc_pcu_dict)
- self.host = self.pcu_name()
-
- def pcu_name(self):
- if self.hostname is not None and self.hostname is not "":
- return self.hostname
- elif self.ip is not None and self.ip is not "":
- return self.ip
- else:
- return None
-
- def nodeidToPort(self, node_id):
- if node_id in self.node_ids:
- for i in range(0, len(self.node_ids)):
- if node_id == self.node_ids[i]:
- return self.ports[i]
-
- raise Exception("No such Node ID: %d" % node_id)
-
-# This class captures the observed pcu records from FindBadPCUs.py
-class PCURecord:
- def __init__(self, pcu_record_dict):
- for field in ['port_status',
- 'dns_status',
- 'entry_complete', ]:
- if field in pcu_record_dict:
- if field == "reboot":
- self.__setattr__("reboot_str", pcu_record_dict[field])
- else:
- self.__setattr__(field, pcu_record_dict[field])
- #else:
- # raise Exception("No such field %s in pcu record dict" % field)
-
-class Transport:
- TELNET = "telnet"
- SSH = "ssh"
- HTTP = "http"
- HTTPS = "https"
- IPAL = "ipal"
- DRAC = "drac"
- AMT = "amt"
-
- TELNET_TIMEOUT = 120
-
- porttypemap = {
- 5869 : DRAC,
- 22 : SSH,
- 23 : TELNET,
- 443 : HTTPS,
- 80 : HTTP,
- 9100 : IPAL,
- 16992 : AMT,
- }
-
- def __init__(self, type, verbose):
- self.type = type
- self.verbose = verbose
- self.transport = None
-
- def open(self, host, username=None, password=None, prompt="User Name"):
- transport = None
-
- if self.type == self.TELNET:
- transport = telnetlib.Telnet(host, timeout=self.TELNET_TIMEOUT)
- transport.set_debuglevel(self.verbose)
- if username is not None:
- self.transport = transport
- self.ifThenSend(prompt, username, ExceptionUsername)
-
- elif self.type == self.SSH:
- if username is not None:
- transport = pyssh.Ssh(username, host)
- transport.set_debuglevel(self.verbose)
- transport.open()
- # TODO: have an ssh set_debuglevel() also...
- else:
- raise Exception("Username cannot be None for ssh transport.")
- elif self.type == self.HTTP:
- # NOTE: this does not work for all web-based services...
- self.url = "http://%s:%d/" % (host,80)
- uri = "%s:%d" % (host,80)
-
- # create authinfo
- authinfo = urllib2.HTTPPasswordMgrWithDefaultRealm()
- authinfo.add_password (None, uri, username, password)
- authhandler = urllib2.HTTPBasicAuthHandler( authinfo )
-
- transport = urllib2.build_opener(authhandler)
- else:
- raise Exception("Unknown transport type: %s" % self.type)
-
- self.transport = transport
- return True
-
- def close(self):
- if self.type == self.TELNET:
- self.transport.close()
- elif self.type == self.SSH:
- self.transport.close()
- elif self.type == self.HTTP:
- pass
- else:
- raise Exception("Unknown transport type %s" % self.type)
- self.transport = None
-
- def write(self, msg):
- return self.send(msg)
-
- def send(self, msg):
- if self.transport == None:
- raise ExceptionNoTransport("transport object is type None")
-
- return self.transport.write(msg)
-
- def sendPassword(self, password, prompt=None):
- if self.type == self.TELNET:
- if prompt == None:
- self.ifThenSend("Password", password, ExceptionPassword)
- else:
- self.ifThenSend(prompt, password, ExceptionPassword)
- elif self.type == self.SSH:
- self.ifThenSend("password:", password, ExceptionPassword)
- elif self.type == self.HTTP:
- pass
- else:
- raise Exception("Unknown transport type: %s" % self.type)
-
- def sendHTTP(self, resource, data):
- if self.verbose:
- print "POSTing '%s' to %s" % (data,self.url + resource)
-
- try:
- f = self.transport.open(self.url + resource ,data)
- r = f.read()
- if self.verbose:
- print r
-
- except urllib2.URLError,err:
- logger.info('Could not open http connection', err)
- return "http transport error"
-
- return 0
-
- def ifThenSend(self, expected, buffer, ErrorClass=ExceptionPrompt):
-
- if self.transport != None:
- output = self.transport.read_until(expected, self.TELNET_TIMEOUT)
- if output.find(expected) == -1:
- print "OUTPUT: --%s--" % output
- raise ErrorClass, "'%s' not found" % expected
- else:
- self.transport.write(buffer + "\r\n")
- else:
- raise ExceptionNoTransport("transport object is type None")
-
- def ifElse(self, expected, ErrorClass):
- try:
- self.transport.read_until(expected, self.TELNET_TIMEOUT)
- except:
- raise ErrorClass("Could not find '%s' within timeout" % expected)
-
-class PCUControl(PCUModel,PCURecord):
-
- """
- There are three cases:
- 1) the pcu_record passed below includes port_status from an
- external probe.
- 2) the external probe failed, and the values are empty
- 3) this call is made independent of port_status.
-
- In the first case, the first open port is used.
- In the third case, the ports are tried in sequence.
-
- In this way, the port_status value serves only as an optimization,
- because closed ports are avoided. The supported_ports value should
- order ports by their preferred usage.
- """
-
- supported_ports = []
-
- def __init__(self, plc_pcu_record, verbose, ignored=None):
- PCUModel.__init__(self, plc_pcu_record)
- PCURecord.__init__(self, plc_pcu_record)
-
- def reboot(self, node_port, dryrun):
-
- port_list = []
- # There are two sources of potential ports. Those that are open and
- # those that are part of the PCU's supported_ports.
- # I think we should start with supported_ports and then filter that
- # by the open ports.
-
- port_list = self.supported_ports
-
- if hasattr(self, 'port_status') and self.port_status:
- # get out the open ports
- port_list = filter(lambda x: self.port_status[x] == "open" , self.port_status.keys())
- port_list = [ int(x) for x in port_list ]
- # take only the open ports that are supported_ports
- port_list = filter(lambda x: x in self.supported_ports, port_list)
- if port_list == []:
- raise ExceptionPort("No Open Port: No transport from open ports")
-
- print port_list
-
- ret = "No implementation for open ports on selected PCU model"
- for port in port_list:
- if port not in Transport.porttypemap:
- continue
-
- type = Transport.porttypemap[port]
- self.transport = Transport(type, verbose)
-
- print "checking for run_%s" % type
- if hasattr(self, "run_%s" % type):
- print "found run_%s" % type
- fxn = getattr(self, "run_%s" % type)
- ret = self.catcherror(fxn, node_port, dryrun)
- if ret == 0: # NOTE: success!, so stop
- break
- else:
- continue
-
- return ret
-
- def run(self, node_port, dryrun):
- """ This function is to be defined by the specific PCU instance. """
- raise Exception("This function is not implemented")
- pass
-
- #def reboot(self, node_port, dryrun):
-
- def catcherror(self, function, node_port, dryrun):
- try:
- return function(node_port, dryrun)
- except ExceptionNotFound, err:
- return "error: " + str(err)
- except ExceptionPassword, err:
- return "Password exception: " + str(err)
- except ExceptionTimeout, err:
- return "Timeout exception: " + str(err)
- except ExceptionUsername, err:
- return "No username prompt: " + str(err)
- except ExceptionSequence, err:
- return "Sequence error: " + str(err)
- except ExceptionPrompt, err:
- return "Prompt exception: " + str(err)
- except ExceptionNoTransport, err:
- return "No Transport: " + str(err)
- except ExceptionPort, err:
- return "No ports exception: " + str(err)
- except socket.error, err:
- return "socket error: timeout: " + str(err)
- except urllib2.HTTPError, err:
- return "HTTPError: " + str(err)
- except urllib2.URLError, err:
- return "URLError: " + str(err)
- except EOFError, err:
- self.transport.close()
- import traceback
- traceback.print_exc()
- return "EOF connection reset" + str(err)
- except Exception, err:
- from monitor.common import email_exception
- email_exception(self.host)
- raise Exception(err)
-
-from pcucontrol.models import *
-
-def pcu_name(pcu):
- if pcu['hostname'] is not None and pcu['hostname'] is not "":
- return pcu['hostname']
- elif pcu['ip'] is not None and pcu['ip'] is not "":
- return pcu['ip']
- else:
- return None
-
def get_pcu_values(pcu_id):
from monitor.database.info.model import FindbadPCURecord
print "pcuid: %s" % pcu_id
@@ -414,129 +108,6 @@ def reboot_policy(nodename, continue_probe, dryrun):
print "return true"
return True
-class Unknown(PCUControl):
- supported_ports = [22,23,80,443,5869,9100,16992]
-
-def model_to_object(modelname):
- if modelname is None:
- return ManualPCU
- if "AMT" in modelname:
- return IntelAMT
- elif "BayTech" in modelname:
- return BayTech
- elif "HPiLO" in modelname:
- return HPiLO
- elif "IPAL" in modelname:
- return IPAL
- elif "APC" in modelname:
- return APCControl
- elif "DRAC" in modelname:
- return DRAC
- elif "WTI" in modelname:
- return WTIIPS4
- elif "ePowerSwitch" in modelname:
- return ePowerSwitchNew
- elif "IPMI" in modelname:
- return IPMI
- elif "BlackBoxPSMaverick" in modelname:
- return BlackBoxPSMaverick
- elif "PM211MIP" in modelname:
- return PM211MIP
- elif "ManualPCU" in modelname:
- return ManualPCU
- else:
- print "UNKNOWN model %s"%modelname
- return Unknown
-
-def reboot_api(node, pcu): #, verbose, dryrun):
- rb_ret = ""
-
- try:
- modelname = pcu['model']
- if modelname:
- # get object instance
- instance = eval('%s(pcu, verbose)' % modelname)
- # get pcu port
- i = pcu['node_ids'].index(node['node_id'])
- p = pcu['ports'][i]
- # reboot
- rb_ret = instance.reboot(p, False)
- else:
- rb_ret = "No modelname in PCU record."
- # TODO: how to handle the weird, georgetown pcus, the drac faults, and ilo faults
- except Exception, err:
- rb_ret = str(err)
-
- return rb_ret
-
-def convert_oldmodelname_to_newmodelname(oldmodelname, pcu_id):
- newmodelname = None
- update = { 'AP79xx' : 'APCControl13p13',
- 'Masterswitch' : 'APCControl13p13',
- 'DS4-RPC' : 'BayTech',
- 'IP-41x_IP-81x' : 'IPAL',
- 'DRAC3' : 'DRAC',
- 'DRAC4' : 'DRAC',
- 'ePowerSwitch' : 'ePowerSwitchOld',
- 'ilo2' : 'HPiLO',
- 'ilo1' : 'HPiLO',
- 'PM211-MIP' : 'PM211MIP',
- 'AMT2.5' : 'IntelAMT',
- 'AMT3.0' : 'IntelAMT',
- 'WTI_IPS-4' : 'WTIIPS4',
- 'unknown' : 'ManualPCU',
- 'DRAC5' : 'DRAC',
- 'ipmi' : 'OpenIPMI',
- 'bbsemaverick' : 'BlackBoxPSMaverick',
- 'manualadmin' : 'ManualPCU',
- }
-
- if oldmodelname in update:
- newmodelname = update[oldmodelname]
- else:
- newmodelname = oldmodelname
-
- if pcu_id in [1102,1163,1055,1111,1231,1113,1127,1128,1148]:
- newmodelname = 'APCControl12p3'
- elif pcu_id in [1110,86]:
- newmodelname = 'APCControl1p4'
- elif pcu_id in [1221,1225,1220,1192]:
- newmodelname = 'APCControl121p3'
- elif pcu_id in [1173,1240,47,1363,1405,1401,1372,1371]:
- newmodelname = 'APCControl121p1'
- elif pcu_id in [1056,1237,1052,1209,1002,1008,1013,1022]:
- newmodelname = 'BayTechCtrlC'
- elif pcu_id in [93]:
- newmodelname = 'BayTechRPC3NC'
- elif pcu_id in [1057]:
- newmodelname = 'BayTechCtrlCUnibe'
- elif pcu_id in [1012]:
- newmodelname = 'BayTechRPC16'
- elif pcu_id in [1089, 1071, 1046, 1035, 1118]:
- newmodelname = 'ePowerSwitchNew'
-
- return newmodelname
-
-def reboot_test_new(nodename, values, verbose, dryrun):
- rb_ret = ""
- if 'plc_pcu_stats' in values:
- values.update(values['plc_pcu_stats'])
-
- try:
- modelname = convert_oldmodelname_to_newmodelname(values['model'], values['pcu_id'])
- if modelname:
- object = eval('%s(values, verbose)' % modelname)
- rb_ret = object.reboot(values[nodename], dryrun)
- else:
- rb_ret = "Not_Run"
- # TODO: how to handle the weird, georgetown pcus, the drac faults, and ilo faults
- except ExceptionPort, err:
- rb_ret = str(err)
- except NameError, err:
- rb_ret = str(err)
-
- return rb_ret
-
def main():
logger.setLevel(logging.DEBUG)
ch = logging.StreamHandler()
diff --git a/monitor/util/command.py b/monitor/util/command.py
index e3e81ca..e5663c3 100644
--- a/monitor/util/command.py
+++ b/monitor/util/command.py
@@ -4,10 +4,12 @@ import subprocess
import signal
import time
import traceback
+import fcntl
DEBUG= 0
class ExceptionTimeout(Exception): pass
+class ExceptionReadTimeout(Exception): pass
COMMAND_TIMEOUT = 60
ssh_options = { 'StrictHostKeyChecking':'no',
'BatchMode':'yes',
@@ -15,15 +17,47 @@ ssh_options = { 'StrictHostKeyChecking':'no',
'ConnectTimeout':'%s' % COMMAND_TIMEOUT}
class Sopen(subprocess.Popen):
- def kill(self, signal = signal.SIGTERM):
- os.kill(self.pid, signal)
+ def kill(self, sig = signal.SIGTERM):
+ try:
+ # NOTE: this also kills parent... so doesn't work like I want.
+ # NOTE: adding 'exec' before the cmd removes the extra sh, and
+ # partially addresses this problem.
+ #os.killpg(os.getpgid(self.pid), signal.SIGKILL)
+ os.kill(self.pid, sig)
+ except OSError:
+ # no such process, due to it already exiting...
+ pass
+
+
+def read_t(stream, count=1, timeout=COMMAND_TIMEOUT*2):
+ if count == 1:
+ retstr = ""
+
+ while True:
+ lin, lout, lerr = select([stream], [], [], timeout)
+ if len(lin) == 0:
+ print "timeout!"
+ raise ExceptionReadTimeout("TIMEOUT reading from command")
-def read_t(stream, count, timeout=COMMAND_TIMEOUT*2):
- lin, lout, lerr = select([stream], [], [], timeout)
- if len(lin) == 0:
- raise ExceptionTimeout("TIMEOUT Running: %s" % cmd)
+ try:
+ outbytes = stream.read(count)
+ except IOError, err:
+ print 'no content yet.'
+ # due to no content.
+ # the select timeout should catch this.
+ continue
- return stream.read(count)
+ if not outbytes:
+ break
+ retstr += outbytes
+
+ return retstr
+ else:
+ lin, lout, lerr = select([stream], [], [], timeout)
+ if len(lin) == 0:
+ raise ExceptionReadTimeout("TIMEOUT reading from command")
+
+ return stream.read(count)
class CMD:
def __init__(self):
@@ -31,15 +65,18 @@ class CMD:
def run_noexcept(self, cmd, timeout=COMMAND_TIMEOUT*2):
- #print "CMD.run_noexcept(%s)" % cmd
try:
return CMD.run(self,cmd,timeout)
except ExceptionTimeout:
print traceback.print_exc()
- return ("", "SCRIPTTIMEOUT")
- except:
+ return ("", "ScriptTimeout")
+ except ExceptionReadTimeout:
+ print traceback.print_exc()
+ return ("", "RunningScriptTimeout")
+ except Exception, err:
from monitor.common import email_exception
email_exception()
+ return ("", str(err))
def system(self, cmd, timeout=COMMAND_TIMEOUT*2):
(o,e) = self.run(cmd, timeout)
@@ -51,16 +88,13 @@ class CMD:
def run(self, cmd, timeout=COMMAND_TIMEOUT*2):
- #print "CMD.run(%s)" % cmd
s = Sopen(cmd, shell=True, stdin=subprocess.PIPE, stdout=subprocess.PIPE, stderr=subprocess.PIPE, close_fds=True)
self.s = s
(f_in, f_out, f_err) = (s.stdin, s.stdout, s.stderr)
- #print "calling select(%s)" % timeout
lout, lin, lerr = select([f_out], [], [f_err], timeout)
- #print "TIMEOUT!!!!!!!!!!!!!!!!!!!"
if len(lin) == 0 and len(lout) == 0 and len(lerr) == 0:
# Reached a timeout! Nuke process so it does not hang.
- #print "KILLING"
+ print "TIMEOUT!!!!!!!!!!!!!!!!!!!"
s.kill(signal.SIGKILL)
raise ExceptionTimeout("TIMEOUT Running: %s" % cmd)
else:
@@ -71,28 +105,26 @@ class CMD:
o_value = ""
e_value = ""
- o_value = f_out.read()
+ #o_value = f_out.read()
+ flags = fcntl.fcntl(f_out, fcntl.F_GETFL)
+ fcntl.fcntl(f_out, fcntl.F_SETFL, flags | os.O_NONBLOCK)
+
+ try:
+ o_value = read_t(f_out,1,30)
+ except ExceptionReadTimeout:
+ s.kill(signal.SIGKILL)
+ raise ExceptionReadTimeout("TIMEOUT: failed to read from cmd: %s" % cmd)
+
e_value = f_err.read()
- #print "striping output"
o_value = o_value.strip()
e_value = e_value.strip()
- #print "OUTPUT -%s-%s-" % (o_value, e_value)
-
- #print "closing files"
f_out.close()
f_in.close()
f_err.close()
- try:
- #print "s.kill()"
- s.kill()
- #print "after s.kill()"
- except OSError:
- # no such process, due to it already exiting...
- pass
+ s.kill(signal.SIGKILL)
- #print o_value, e_value
return (o_value, e_value)
def runargs(self, args, timeout=COMMAND_TIMEOUT*2):
@@ -117,11 +149,7 @@ class CMD:
f_out.close()
f_in.close()
f_err.close()
- try:
- s.kill()
- except OSError:
- # no such process, due to it already exiting...
- pass
+ s.kill(signal.SIGKILL)
return (o_value, e_value)
@@ -164,17 +192,9 @@ class SSH(CMD):
return CMD.run_noexcept(self, cmd)
def run_noexcept2(self, cmd, timeout=COMMAND_TIMEOUT*2):
- cmd = "ssh -p %s %s %s@%s %s" % (self.port, self.__options_to_str(),
+ cmd = "exec ssh -p %s %s %s@%s %s" % (self.port, self.__options_to_str(),
self.user, self.host, cmd)
- #print "SSH.run_noexcept2(%s)" % cmd
r = CMD.run_noexcept(self, cmd, timeout)
-
- # XXX: this may be resulting in deadlocks... not sure.
- #if self.s.returncode is None:
- # #self.s.kill()
- # self.s.kill(signal.SIGKILL)
- # self.s.wait()
- # self.ret = self.s.returncode
self.ret = -1
return r
diff --git a/nodeinfo.py b/nodeinfo.py
index 9afed5c..3248707 100755
--- a/nodeinfo.py
+++ b/nodeinfo.py
@@ -8,7 +8,7 @@ from monitor import util
from monitor import parser as parsermodule
from monitor import database
-from pcucontrol import reboot
+from monitor import reboot
import time
from monitor.model import *
diff --git a/nodequery.py b/nodequery.py
index 781e841..7a53df0 100755
--- a/nodequery.py
+++ b/nodequery.py
@@ -13,7 +13,6 @@ import time
import re
import string
-from pcucontrol import reboot
from monitor.wrapper import plc, plccache
api = plc.getAuthAPI()
@@ -383,8 +382,6 @@ def main():
#fbnodes = FindbadNodeRecord.select(FindbadNodeRecord.q.hostname, orderBy='date_checked',distinct=True).reversed()
fb = None
- #reboot.fb = fbpcu
-
if config.nodelist:
nodelist = util.file.getListFromFile(config.nodelist)
else:
diff --git a/pcubad.py b/pcubad.py
index 181f001..13fce72 100755
--- a/pcubad.py
+++ b/pcubad.py
@@ -7,7 +7,7 @@ import time
from datetime import datetime,timedelta
from monitor import database
-from pcucontrol import reboot
+from monitor import reboot
from monitor import parser as parsermodule
from monitor import config
from monitor.database.info.model import HistoryPCURecord, FindbadPCURecord
diff --git a/pcucontrol/reboot.py b/pcucontrol/reboot.py
index b762d89..a0ba1f2 100755
--- a/pcucontrol/reboot.py
+++ b/pcucontrol/reboot.py
@@ -11,13 +11,12 @@ import urllib2
import urllib
import threading, popen2
import array, struct
-from monitor.wrapper import plc
import base64
from subprocess import PIPE, Popen
import pcucontrol.transports.ssh.pxssh as pxssh
import pcucontrol.transports.ssh.pexpect as pexpect
import socket
-from monitor.util import command
+
# Use our versions of telnetlib and pyssh
@@ -25,8 +24,6 @@ sys.path.insert(0, os.path.dirname(sys.argv[0]))
import pcucontrol.transports.telnetlib as telnetlib
sys.path.insert(0, os.path.dirname(sys.argv[0]) + "/pyssh")
import pcucontrol.transports.pyssh as pyssh
-from monitor import config
-
# Event class ID from pcu events
#NODE_POWER_CONTROL = 3
@@ -35,7 +32,6 @@ from monitor import config
#MONITOR_USER_ID = 11142
import logging
-logger = logging.getLogger("monitor")
verbose = 1
#dryrun = 0;
@@ -206,7 +202,7 @@ class Transport:
print r
except urllib2.URLError,err:
- logger.info('Could not open http connection', err)
+ print 'Could not open http connection', err
return "http transport error"
return 0
@@ -345,73 +341,6 @@ def pcu_name(pcu):
else:
return None
-def get_pcu_values(pcu_id):
- from monitor.database.info.model import FindbadPCURecord
- print "pcuid: %s" % pcu_id
- try:
- pcurec = FindbadPCURecord.get_latest_by(plc_pcuid=pcu_id).first()
- if pcurec:
- values = pcurec.to_dict()
- else:
- values = None
- except:
- values = None
-
- return values
-
-def reboot(nodename):
- return reboot_policy(nodename, True, False)
-
-def reboot_str(nodename):
- global verbose
- continue_probe = True
- dryrun=False
-
- pcu = plc.getpcu(nodename)
- if not pcu:
- logger.debug("no pcu for %s" % nodename)
- print "no pcu for %s" % nodename
- return False # "%s has no pcu" % nodename
-
- values = get_pcu_values(pcu['pcu_id'])
- if values == None:
- logger.debug("No values for pcu probe %s" % nodename)
- print "No values for pcu probe %s" % nodename
- return False #"no info for pcu_id %s" % pcu['pcu_id']
-
- # Try the PCU first
- logger.debug("Trying PCU %s %s" % (pcu['hostname'], pcu['model']))
-
- ret = reboot_test_new(nodename, values, verbose, dryrun)
- return ret
-
-def reboot_policy(nodename, continue_probe, dryrun):
- global verbose
-
- pcu = plc.getpcu(nodename)
- if not pcu:
- logger.debug("no pcu for %s" % nodename)
- print "no pcu for %s" % nodename
- return False # "%s has no pcu" % nodename
-
- values = get_pcu_values(pcu['pcu_id'])
- if values == None:
- logger.debug("No values for pcu probe %s" % nodename)
- print "No values for pcu probe %s" % nodename
- return False #"no info for pcu_id %s" % pcu['pcu_id']
-
- # Try the PCU first
- logger.debug("Trying PCU %s %s" % (pcu['hostname'], pcu['model']))
-
- ret = reboot_test_new(nodename, values, verbose, dryrun)
-
- if ret != 0:
- print ret
- return False
- else:
- print "return true"
- return True
-
class Unknown(PCUControl):
supported_ports = [22,23,80,443,5869,9100,16992]
@@ -446,7 +375,7 @@ def model_to_object(modelname):
print "UNKNOWN model %s"%modelname
return Unknown
-def reboot_api(node, pcu): #, verbose, dryrun):
+def reboot_api(node, pcu):
rb_ret = ""
try:
@@ -536,36 +465,7 @@ def reboot_test_new(nodename, values, verbose, dryrun):
return rb_ret
def main():
- logger.setLevel(logging.DEBUG)
- ch = logging.StreamHandler()
- ch.setLevel(logging.DEBUG)
- formatter = logging.Formatter('LOGGER - %(message)s')
- ch.setFormatter(formatter)
- logger.addHandler(ch)
-
- try:
- if "test" in sys.argv:
- dryrun = True
- else:
- dryrun = False
-
- for node in sys.argv[1:]:
- if node == "test": continue
-
- print "Rebooting %s" % node
- if reboot_policy(node, True, dryrun):
- print "success"
- else:
- print "failed"
- except Exception, err:
- import traceback; traceback.print_exc()
- from monitor.common import email_exception
- email_exception(node)
- print err
+ print "this does not work."
if __name__ == '__main__':
- logger = logging.getLogger("monitor")
main()
- f = open("/tmp/rebootlog", 'a')
- f.write("reboot %s\n" % sys.argv)
- f.close()
diff --git a/sitebad.py b/sitebad.py
index f8524f0..15c8f1d 100755
--- a/sitebad.py
+++ b/sitebad.py
@@ -7,7 +7,6 @@ import time
from datetime import datetime,timedelta
from monitor import database
-from pcucontrol import reboot
from monitor import parser as parsermodule
from monitor import config
from monitor.database.info.model import HistorySiteRecord, FindbadNodeRecord, session
diff --git a/siteinfo.py b/siteinfo.py
index cfce458..6fe6496 100755
--- a/siteinfo.py
+++ b/siteinfo.py
@@ -4,7 +4,6 @@ from monitor.wrapper import plc
api = plc.getAuthAPI()
from monitor import database
-from pcucontrol import reboot
import time
from monitor.common import *
--
2.47.0