2 # NEPI, a framework to manage network experiments
3 # Copyright (C) 2013 INRIA
5 # This program is free software: you can redistribute it and/or modify
6 # it under the terms of the GNU General Public License as published by
7 # the Free Software Foundation, either version 3 of the License, or
8 # (at your option) any later version.
10 # This program is distributed in the hope that it will be useful,
11 # but WITHOUT ANY WARRANTY; without even the implied warranty of
12 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
13 # GNU General Public License for more details.
15 # You should have received a copy of the GNU General Public License
16 # along with this program. If not, see <http://www.gnu.org/licenses/>.
18 # Author: Alina Quereilhac <alina.quereilhac@inria.fr>
20 from nepi.execution.attribute import Attribute, Flags, Types
21 from nepi.execution.resource import ResourceManager, clsinit_copy, \
23 from nepi.resources.linux import rpmfuncs, debfuncs
24 from nepi.util import sshfuncs, execfuncs
25 from nepi.util.sshfuncs import ProcStatus
36 # TODO: Unify delays!!
37 # TODO: Validate outcome of uploads!!
41 Error codes that the rexitcode function can return if unable to
42 check the exit code of a spawned process
51 Supported flavors of Linux OS
56 FEDORA_8 = 1 << 3 | FEDORA
57 FEDORA_12 = 1 << 4 | FEDORA
58 FEDORA_14 = 1 << 5 | FEDORA
61 class LinuxNode(ResourceManager):
63 .. class:: Class Args :
65 :param ec: The Experiment controller
66 :type ec: ExperimentController
67 :param guid: guid of the RM
72 There are different ways in which commands can be executed using the
73 LinuxNode interface (i.e. 'execute' - blocking and non blocking, 'run',
78 * 'execute' (blocking mode) :
80 HOW IT WORKS: 'execute', forks a process and run the
81 command, synchronously, attached to the terminal, in
83 The execute method will block until the command returns
84 the result on 'out', 'err' (so until it finishes executing).
86 USAGE: short-lived commands that must be executed attached
87 to a terminal and in foreground, for which it IS necessary
88 to block until the command has finished (e.g. if you want
89 to run 'ls' or 'cat').
91 * 'execute' (NON blocking mode - blocking = False) :
93 HOW IT WORKS: Same as before, except that execute method
94 will return immediately (even if command still running).
96 USAGE: long-lived commands that must be executed attached
97 to a terminal and in foreground, but for which it is not
98 necessary to block until the command has finished. (e.g.
99 start an application using X11 forwarding)
103 HOW IT WORKS: Connects to the host ( using SSH if remote)
104 and launches the command in background, detached from any
105 terminal (daemonized), and returns. The command continues to
106 run remotely, but since it is detached from the terminal,
107 its pipes (stdin, stdout, stderr) can't be redirected to the
108 console (as normal non detached processes would), and so they
109 are explicitly redirected to files. The pidfile is created as
110 part of the process of launching the command. The pidfile
111 holds the pid and ppid of the process forked in background,
112 so later on it is possible to check whether the command is still
115 USAGE: long-lived commands that can run detached in background,
116 for which it is NOT necessary to block (wait) until the command
117 has finished. (e.g. start an application that is not using X11
118 forwarding. It can run detached and remotely in background)
122 HOW IT WORKS: Similar to 'run' except that it 'blocks' until
123 the command has finished execution. It also checks whether
124 errors occurred during runtime by reading the exitcode file,
125 which contains the exit code of the command that was run
126 (checking stderr only is not always reliable since many
127 commands throw debugging info to stderr and the only way to
128 automatically know whether an error really happened is to
129 check the process exit code).
131 Another difference with respect to 'run', is that instead
132 of directly executing the command as a bash command line,
133 it uploads the command to a bash script and runs the script.
134 This allows to use the bash script to debug errors, since
135 it remains at the remote host and can be run manually to
138 USAGE: medium-lived commands that can run detached in
139 background, for which it IS necessary to block (wait) until
140 the command has finished. (e.g. Package installation,
141 source compilation, file download, etc)
144 _rtype = "linux::Node"
145 _help = "Controls Linux host machines ( either localhost or a host " \
146 "that can be accessed using a SSH key)"
150 def _register_attributes(cls):
151 hostname = Attribute("hostname", "Hostname of the machine",
152 flags = Flags.Design)
154 username = Attribute("username", "Local account username",
155 flags = Flags.Credential)
157 port = Attribute("port", "SSH port", flags = Flags.Design)
159 home = Attribute("home",
160 "Experiment home directory to store all experiment related files",
161 flags = Flags.Design)
163 identity = Attribute("identity", "SSH identity file",
164 flags = Flags.Credential)
166 server_key = Attribute("serverKey", "Server public key",
167 flags = Flags.Design)
169 clean_home = Attribute("cleanHome", "Remove all nepi files and directories "
170 " from node home folder before starting experiment",
173 flags = Flags.Design)
175 clean_experiment = Attribute("cleanExperiment", "Remove all files and directories "
176 " from a previous same experiment, before the new experiment starts",
179 flags = Flags.Design)
181 clean_processes = Attribute("cleanProcesses",
182 "Kill all running processes before starting experiment",
185 flags = Flags.Design)
187 tear_down = Attribute("tearDown", "Bash script to be executed before " + \
188 "releasing the resource",
189 flags = Flags.Design)
191 gateway_user = Attribute("gatewayUser", "Gateway account username",
192 flags = Flags.Design)
194 gateway = Attribute("gateway", "Hostname of the gateway machine",
195 flags = Flags.Design)
197 ip = Attribute("ip", "Linux host public IP address. "
198 "Must not be modified by the user unless hostname is 'localhost'",
199 flags = Flags.Design)
201 cls._register_attribute(hostname)
202 cls._register_attribute(username)
203 cls._register_attribute(port)
204 cls._register_attribute(home)
205 cls._register_attribute(identity)
206 cls._register_attribute(server_key)
207 cls._register_attribute(clean_home)
208 cls._register_attribute(clean_experiment)
209 cls._register_attribute(clean_processes)
210 cls._register_attribute(tear_down)
211 cls._register_attribute(gateway_user)
212 cls._register_attribute(gateway)
213 cls._register_attribute(ip)
215 def __init__(self, ec, guid):
216 super(LinuxNode, self).__init__(ec, guid)
218 # home directory at Linux host
221 # lock to prevent concurrent applications on the same node,
222 # to execute commands at the same time. There are potential
223 # concurrency issues when using SSH to a same host from
224 # multiple threads. There are also possible operational
225 # issues, e.g. an application querying the existence
226 # of a file or folder prior to its creation, and another
227 # application creating the same file or folder in between.
228 self._node_lock = threading.Lock()
230 def log_message(self, msg):
231 return " guid %d - host %s - %s " % (self.guid,
232 self.get("hostname"), msg)
236 home = self.get("home") or ""
237 if not home.startswith("/"):
238 home = os.path.join(self._home_dir, home)
243 return os.path.join(self.home_dir, ".nepi")
247 return os.path.join(self.nepi_home, "nepi-usr")
251 return os.path.join(self.usr_dir, "lib")
255 return os.path.join(self.usr_dir, "bin")
259 return os.path.join(self.usr_dir, "src")
263 return os.path.join(self.usr_dir, "share")
267 return os.path.join(self.nepi_home, "nepi-exp")
271 return os.path.join(self.exp_dir, self.ec.exp_id)
275 return os.path.join(self.exp_home, "node-%d" % self.guid)
279 return os.path.join(self.node_home, self.ec.run_id)
286 if not self.localhost and not self.get("username"):
287 msg = "Can't resolve OS, insufficient data "
289 raise RuntimeError, msg
293 if out.find("Debian") == 0:
294 self._os = OSType.DEBIAN
295 elif out.find("Ubuntu") ==0:
296 self._os = OSType.UBUNTU
297 elif out.find("Fedora release") == 0:
298 self._os = OSType.FEDORA
299 if out.find("Fedora release 8") == 0:
300 self._os = OSType.FEDORA_8
301 elif out.find("Fedora release 12") == 0:
302 self._os = OSType.FEDORA_12
303 elif out.find("Fedora release 14") == 0:
304 self._os = OSType.FEDORA_14
306 msg = "Unsupported OS"
308 raise RuntimeError, "%s - %s " %( msg, out )
313 # The underlying SSH layer will sometimes return an empty
314 # output (even if the command was executed without errors).
315 # To work arround this, repeat the operation N times or
316 # until the result is not empty string
319 (out, err), proc = self.execute("cat /etc/issue",
323 trace = traceback.format_exc()
324 msg = "Error detecting OS: %s " % trace
325 self.error(msg, out, err)
331 return (self.os & (OSType.DEBIAN|OSType.UBUNTU))
335 return (self.os & OSType.FEDORA)
339 return self.get("hostname") in ['localhost', '127.0.0.1', '::1']
341 def do_provision(self):
342 # check if host is alive
343 if not self.is_alive():
344 msg = "Deploy failed. Unresponsive node %s" % self.get("hostname")
346 raise RuntimeError, msg
350 if self.get("cleanProcesses"):
351 self.clean_processes()
353 if self.get("cleanHome"):
356 if self.get("cleanExperiment"):
357 self.clean_experiment()
359 # Create shared directory structure and node home directory
360 paths = [self.lib_dir,
368 # Get Public IP address if possible
369 if not self.get("ip"):
371 ip = sshfuncs.gethostbyname(self.get("hostname"))
373 msg = "DNS can not resolve hostname %s" % self.get("hostname")
378 super(LinuxNode, self).do_provision()
381 if self.state == ResourceState.NEW:
382 self.info("Deploying node")
386 # Node needs to wait until all associated interfaces are
387 # ready before it can finalize deployment
388 from nepi.resources.linux.interface import LinuxInterface
389 ifaces = self.get_connected(LinuxInterface.get_rtype())
391 if iface.state < ResourceState.READY:
392 self.ec.schedule(self.reschedule_delay, self.deploy)
395 super(LinuxNode, self).do_deploy()
397 def do_release(self):
398 rms = self.get_connected()
400 # Node needs to wait until all associated RMs are released
401 # before it can be released
402 if rm.state != ResourceState.RELEASED:
403 self.ec.schedule(self.reschedule_delay, self.release)
406 tear_down = self.get("tearDown")
408 self.execute(tear_down)
410 self.clean_processes()
412 super(LinuxNode, self).do_release()
414 def valid_connection(self, guid):
418 def clean_processes(self):
419 self.info("Cleaning up processes")
424 if self.get("username") != 'root':
425 cmd = ("sudo -S killall tcpdump || /bin/true ; " +
426 "sudo -S killall -u %s || /bin/true ; " % self.get("username"))
428 if self.state >= ResourceState.READY:
430 pids = pickle.load(open("/tmp/save.proc", "rb"))
432 ps_aux = "ps aux |awk '{print $2,$11}'"
433 (out, err), proc = self.execute(ps_aux)
435 for line in out.strip().split("\n"):
436 parts = line.strip().split(" ")
437 pids_temp[parts[0]] = parts[1]
438 kill_pids = set(pids_temp.items()) - set(pids.items())
439 kill_pids = ' '.join(dict(kill_pids).keys())
441 cmd = ("killall tcpdump || /bin/true ; " +
442 "kill $(ps aux | grep '[n]epi' | awk '{print $2}') || /bin/true ; " +
443 "kill %s || /bin/true ; " % kill_pids)
445 cmd = ("killall tcpdump || /bin/true ; " +
446 "kill $(ps aux | grep '[n]epi' | awk '{print $2}') || /bin/true ; ")
448 cmd = ("killall tcpdump || /bin/true ; " +
449 "kill $(ps aux | grep '[n]epi' | awk '{print $2}') || /bin/true ; ")
451 (out, err), proc = self.execute(cmd, retry = 1, with_lock = True)
453 def clean_home(self):
454 """ Cleans all NEPI related folders in the Linux host
456 self.info("Cleaning up home")
458 cmd = "cd %s ; find . -maxdepth 1 -name \.nepi -execdir rm -rf {} + " % (
461 return self.execute(cmd, with_lock = True)
463 def clean_experiment(self):
464 """ Cleans all experiment related files in the Linux host.
465 It preserves NEPI files and folders that have a multi experiment
468 self.info("Cleaning up experiment files")
470 cmd = "cd %s ; find . -maxdepth 1 -name '%s' -execdir rm -rf {} + " % (
474 return self.execute(cmd, with_lock = True)
476 def execute(self, command,
482 connect_timeout = 30,
483 strict_host_checking = False,
488 """ Notice that this invocation will block until the
489 execution finishes. If this is not the desired behavior,
490 use 'run' instead."""
493 (out, err), proc = execfuncs.lexec(command,
494 user = self.get("username"), # still problem with localhost
499 # If the execute command is blocking, we don't want to keep
500 # the node lock. This lock is used to avoid race conditions
501 # when creating the ControlMaster sockets. A more elegant
502 # solution is needed.
503 with self._node_lock:
504 (out, err), proc = sshfuncs.rexec(
506 host = self.get("hostname"),
507 user = self.get("username"),
508 port = self.get("port"),
509 gwuser = self.get("gatewayUser"),
510 gw = self.get("gateway"),
513 identity = self.get("identity"),
514 server_key = self.get("serverKey"),
517 forward_x11 = forward_x11,
519 connect_timeout = connect_timeout,
520 persistent = persistent,
522 strict_host_checking = strict_host_checking
525 (out, err), proc = sshfuncs.rexec(
527 host = self.get("hostname"),
528 user = self.get("username"),
529 port = self.get("port"),
530 gwuser = self.get("gatewayUser"),
531 gw = self.get("gateway"),
534 identity = self.get("identity"),
535 server_key = self.get("serverKey"),
538 forward_x11 = forward_x11,
540 connect_timeout = connect_timeout,
541 persistent = persistent,
543 strict_host_checking = strict_host_checking
546 return (out, err), proc
548 def run(self, command, home,
556 strict_host_checking = False):
558 self.debug("Running command '%s'" % command)
561 (out, err), proc = execfuncs.lspawn(command, pidfile,
563 create_home = create_home,
564 stdin = stdin or '/dev/null',
565 stdout = stdout or '/dev/null',
566 stderr = stderr or '/dev/null',
569 with self._node_lock:
570 (out, err), proc = sshfuncs.rspawn(
574 create_home = create_home,
575 stdin = stdin or '/dev/null',
576 stdout = stdout or '/dev/null',
577 stderr = stderr or '/dev/null',
579 host = self.get("hostname"),
580 user = self.get("username"),
581 port = self.get("port"),
582 gwuser = self.get("gatewayUser"),
583 gw = self.get("gateway"),
585 identity = self.get("identity"),
586 server_key = self.get("serverKey"),
588 strict_host_checking = strict_host_checking
591 return (out, err), proc
593 def getpid(self, home, pidfile = "pidfile"):
595 pidtuple = execfuncs.lgetpid(os.path.join(home, pidfile))
597 with self._node_lock:
598 pidtuple = sshfuncs.rgetpid(
599 os.path.join(home, pidfile),
600 host = self.get("hostname"),
601 user = self.get("username"),
602 port = self.get("port"),
603 gwuser = self.get("gatewayUser"),
604 gw = self.get("gateway"),
606 identity = self.get("identity"),
607 server_key = self.get("serverKey"),
608 strict_host_checking = False
613 def status(self, pid, ppid):
615 status = execfuncs.lstatus(pid, ppid)
617 with self._node_lock:
618 status = sshfuncs.rstatus(
620 host = self.get("hostname"),
621 user = self.get("username"),
622 port = self.get("port"),
623 gwuser = self.get("gatewayUser"),
624 gw = self.get("gateway"),
626 identity = self.get("identity"),
627 server_key = self.get("serverKey"),
628 strict_host_checking = False
633 def kill(self, pid, ppid, sudo = False):
636 status = self.status(pid, ppid)
638 if status == sshfuncs.ProcStatus.RUNNING:
640 (out, err), proc = execfuncs.lkill(pid, ppid, sudo)
642 with self._node_lock:
643 (out, err), proc = sshfuncs.rkill(
645 host = self.get("hostname"),
646 user = self.get("username"),
647 port = self.get("port"),
648 gwuser = self.get("gatewayUser"),
649 gw = self.get("gateway"),
652 identity = self.get("identity"),
653 server_key = self.get("serverKey"),
654 strict_host_checking = False
657 return (out, err), proc
659 def copy(self, src, dst):
661 (out, err), proc = execfuncs.lcopy(src, dst,
664 with self._node_lock:
665 (out, err), proc = sshfuncs.rcopy(
667 port = self.get("port"),
668 gwuser = self.get("gatewayUser"),
669 gw = self.get("gateway"),
670 identity = self.get("identity"),
671 server_key = self.get("serverKey"),
673 strict_host_checking = False)
675 return (out, err), proc
677 def upload(self, src, dst, text = False, overwrite = True,
678 raise_on_error = True):
679 """ Copy content to destination
681 src string with the content to copy. Can be:
683 - a string with the path to a local file
684 - a string with a semi-colon separeted list of local files
685 - a string with a local directory
687 dst string with destination path on the remote host (remote is
690 text src is text input, it must be stored into a temp file before
693 # If source is a string input
695 if text and not os.path.isfile(src):
696 # src is text input that should be uploaded as file
697 # create a temporal file with the content to upload
698 f = tempfile.NamedTemporaryFile(delete=False)
703 # If dst files should not be overwritten, check that the files do not
705 if isinstance(src, str):
706 src = map(str.strip, src.split(";"))
708 if overwrite == False:
709 src = self.filter_existing_files(src, dst)
711 return ("", ""), None
713 if not self.localhost:
714 # Build destination as <user>@<server>:<path>
715 dst = "%s@%s:%s" % (self.get("username"), self.get("hostname"), dst)
717 ((out, err), proc) = self.copy(src, dst)
724 msg = " Failed to upload files - src: %s dst: %s" % (";".join(src), dst)
725 self.error(msg, out, err)
727 msg = "%s out: %s err: %s" % (msg, out, err)
729 raise RuntimeError, msg
731 return ((out, err), proc)
733 def download(self, src, dst, raise_on_error = True):
734 if not self.localhost:
735 # Build destination as <user>@<server>:<path>
736 src = "%s@%s:%s" % (self.get("username"), self.get("hostname"), src)
738 ((out, err), proc) = self.copy(src, dst)
741 msg = " Failed to download files - src: %s dst: %s" % (";".join(src), dst)
742 self.error(msg, out, err)
745 raise RuntimeError, msg
747 return ((out, err), proc)
749 def install_packages_command(self, packages):
752 command = rpmfuncs.install_packages_command(self.os, packages)
754 command = debfuncs.install_packages_command(self.os, packages)
756 msg = "Error installing packages ( OS not known ) "
757 self.error(msg, self.os)
758 raise RuntimeError, msg
762 def install_packages(self, packages, home, run_home = None,
763 raise_on_error = True):
764 """ Install packages in the Linux host.
766 'home' is the directory to upload the package installation script.
767 'run_home' is the directory from where to execute the script.
769 command = self.install_packages_command(packages)
771 run_home = run_home or home
773 (out, err), proc = self.run_and_wait(command, run_home,
774 shfile = os.path.join(home, "instpkg.sh"),
775 pidfile = "instpkg_pidfile",
776 ecodefile = "instpkg_exitcode",
777 stdout = "instpkg_stdout",
778 stderr = "instpkg_stderr",
780 raise_on_error = raise_on_error)
782 return (out, err), proc
784 def remove_packages(self, packages, home, run_home = None,
785 raise_on_error = True):
786 """ Uninstall packages from the Linux host.
788 'home' is the directory to upload the package un-installation script.
789 'run_home' is the directory from where to execute the script.
792 command = rpmfuncs.remove_packages_command(self.os, packages)
794 command = debfuncs.remove_packages_command(self.os, packages)
796 msg = "Error removing packages ( OS not known ) "
798 raise RuntimeError, msg
800 run_home = run_home or home
802 (out, err), proc = self.run_and_wait(command, run_home,
803 shfile = os.path.join(home, "rmpkg.sh"),
804 pidfile = "rmpkg_pidfile",
805 ecodefile = "rmpkg_exitcode",
806 stdout = "rmpkg_stdout",
807 stderr = "rmpkg_stderr",
809 raise_on_error = raise_on_error)
811 return (out, err), proc
813 def mkdir(self, paths, clean = False):
814 """ Paths is either a single remote directory path to create,
815 or a list of directories to create.
820 if isinstance(paths, str):
823 cmd = " ; ".join(map(lambda path: "mkdir -p %s" % path, paths))
825 return self.execute(cmd, with_lock = True)
827 def rmdir(self, paths):
828 """ Paths is either a single remote directory path to delete,
829 or a list of directories to delete.
832 if isinstance(paths, str):
835 cmd = " ; ".join(map(lambda path: "rm -rf %s" % path, paths))
837 return self.execute(cmd, with_lock = True)
839 def run_and_wait(self, command, home,
845 ecodefile="exitcode",
851 raise_on_error=True):
853 Uploads the 'command' to a bash script in the host.
854 Then runs the script detached in background in the host, and
855 busy-waites until the script finishes executing.
858 if not shfile.startswith("/"):
859 shfile = os.path.join(home, shfile)
861 self.upload_command(command,
863 ecodefile = ecodefile,
865 overwrite = overwrite)
867 command = "bash %s" % shfile
868 # run command in background in remote host
869 (out, err), proc = self.run(command, home,
877 # check no errors occurred
879 msg = " Failed to run command '%s' " % command
880 self.error(msg, out, err)
882 raise RuntimeError, msg
884 # Wait for pid file to be generated
885 pid, ppid = self.wait_pid(
888 raise_on_error = raise_on_error)
891 # wait until command finishes to execute
892 self.wait_run(pid, ppid)
894 (eout, err), proc = self.check_errors(home,
895 ecodefile = ecodefile,
898 # Out is what was written in the stderr file
900 msg = " Failed to run command '%s' " % command
901 self.error(msg, eout, err)
904 raise RuntimeError, msg
906 (out, oerr), proc = self.check_output(home, stdout)
908 return (out, err), proc
910 def exitcode(self, home, ecodefile = "exitcode"):
912 Get the exit code of an application.
913 Returns an integer value with the exit code
915 (out, err), proc = self.check_output(home, ecodefile)
917 # Succeeded to open file, return exit code in the file
920 return int(out.strip())
922 # Error in the content of the file!
923 return ExitCode.CORRUPTFILE
925 # No such file or directory
926 if proc.returncode == 1:
927 return ExitCode.FILENOTFOUND
929 # Other error from 'cat'
930 return ExitCode.ERROR
932 def upload_command(self, command,
934 ecodefile="exitcode",
937 """ Saves the command as a bash script file in the remote host, and
938 forces to save the exit code of the command execution to the ecodefile
941 if not (command.strip().endswith(";") or command.strip().endswith("&")):
944 # The exit code of the command will be stored in ecodefile
945 command = " { %(command)s } ; echo $? > %(ecodefile)s ;" % {
947 'ecodefile': ecodefile,
951 environ = self.format_environment(env)
953 # Add environ to command
954 command = environ + command
956 return self.upload(command, shfile, text=True, overwrite=overwrite)
958 def format_environment(self, env, inline=False):
959 """ Formats the environment variables for a command to be executed
960 either as an inline command
961 (i.e. export PYTHONPATH=src/..; export LALAL= ..;python script.py) or
962 as a bash script (i.e. export PYTHONPATH=src/.. \n export LALA=.. \n)
964 if not env: return ""
966 # Remove extra white spaces
967 env = re.sub(r'\s+', ' ', env.strip())
969 sep = ";" if inline else "\n"
970 return sep.join(map(lambda e: " export %s" % e, env.split(" "))) + sep
972 def check_errors(self, home,
973 ecodefile = "exitcode",
975 """ Checks whether errors occurred while running a command.
976 It first checks the exit code for the command, and only if the
977 exit code is an error one it returns the error output.
983 # get exit code saved in the 'exitcode' file
984 ecode = self.exitcode(home, ecodefile)
986 if ecode in [ ExitCode.CORRUPTFILE, ExitCode.ERROR ]:
987 err = "Error retrieving exit code status from file %s/%s" % (home, ecodefile)
988 elif ecode > 0 or ecode == ExitCode.FILENOTFOUND:
989 # The process returned an error code or didn't exist.
990 # Check standard error.
991 (err, eerr), proc = self.check_output(home, stderr)
993 # If the stderr file was not found, assume nothing bad happened,
994 # and just ignore the error.
995 # (cat returns 1 for error "No such file or directory")
996 if ecode == ExitCode.FILENOTFOUND and proc.poll() == 1:
999 return ("", err), proc
1001 def wait_pid(self, home, pidfile = "pidfile", raise_on_error = False):
1002 """ Waits until the pid file for the command is generated,
1003 and returns the pid and ppid of the process """
1008 pidtuple = self.getpid(home = home, pidfile = pidfile)
1011 pid, ppid = pidtuple
1017 msg = " Failed to get pid for pidfile %s/%s " % (
1022 raise RuntimeError, msg
1026 def wait_run(self, pid, ppid, trial = 0):
1027 """ wait for a remote process to finish execution """
1031 status = self.status(pid, ppid)
1033 if status is ProcStatus.FINISHED:
1035 elif status is not ProcStatus.RUNNING:
1038 # If it takes more than 20 seconds to start, then
1039 # asume something went wrong
1043 # The app is running, just wait...
1046 def check_output(self, home, filename):
1047 """ Retrives content of file """
1048 (out, err), proc = self.execute("cat %s" %
1049 os.path.join(home, filename), retry = 1, with_lock = True)
1050 return (out, err), proc
1053 """ Checks if host is responsive
1059 msg = "Unresponsive host. Wrong answer. "
1061 # The underlying SSH layer will sometimes return an empty
1062 # output (even if the command was executed without errors).
1063 # To work arround this, repeat the operation N times or
1064 # until the result is not empty string
1066 (out, err), proc = self.execute("echo 'ALIVE'",
1070 if out.find("ALIVE") > -1:
1073 trace = traceback.format_exc()
1074 msg = "Unresponsive host. Error reaching host: %s " % trace
1076 self.error(msg, out, err)
1079 def find_home(self):
1080 """ Retrieves host home directory
1082 # The underlying SSH layer will sometimes return an empty
1083 # output (even if the command was executed without errors).
1084 # To work arround this, repeat the operation N times or
1085 # until the result is not empty string
1086 msg = "Impossible to retrieve HOME directory"
1088 (out, err), proc = self.execute("echo ${HOME}",
1092 if out.strip() != "":
1093 self._home_dir = out.strip()
1095 trace = traceback.format_exc()
1096 msg = "Impossible to retrieve HOME directory %s" % trace
1098 if not self._home_dir:
1100 raise RuntimeError, msg
1102 def filter_existing_files(self, src, dst):
1103 """ Removes files that already exist in the Linux host from src list
1105 # construct a dictionary with { dst: src }
1106 dests = dict(map(lambda s: (os.path.join(dst, os.path.basename(s)), s), src)) \
1107 if len(src) > 1 else dict({dst: src[0]})
1110 for d in dests.keys():
1111 command.append(" [ -f %(dst)s ] && echo '%(dst)s' " % {'dst' : d} )
1113 command = ";".join(command)
1115 (out, err), proc = self.execute(command, retry = 1, with_lock = True)
1117 for d in dests.keys():
1118 if out.find(d) > -1:
1124 return dests.values()