Thread safety in jnpr.junos.Device

[Chris Lorier]

Hi,

I am trying to log into multiple devices, but I am hitting thread safety issues with the Device object.

I am running the following code:

import sys

import getpass

import logging

from concurrent.futures import ThreadPoolExecutor

from jnpr.junos import Device

logger = logging.getLogger(__name__)

class OpenAndPrintHello(object):

    def __init__(self, target):

        self.device = Device(

            target,

            user=getpass.getuser(),

            port=22,

            autoprobe=False

            )

    def __call__(self):

        self.device.open()

        logging.warning("hello")

        self.device.close()

with ThreadPoolExecutor() as executor:

    for target in sys.argv[1:]:

        executor.submit(OpenAndPrintHello(target))

 If I give it more than 11 devices to log into I start to get errors:

Unknown exception: module 'pkg_resources' has no attribute 'iter_entry_points'

Traceback (most recent call last):

  File "/home/chrislorier/envs/netconf/lib/python3.5/site-packages/paramiko/transport.py", line 1783, in run

    self.kex_engine.parse_next(ptype, m)

  File "/home/chrislorier/envs/netconf/lib/python3.5/site-packages/paramiko/kex_group1.py", line 75, in parse_next

    return self._parse_kexdh_reply(m)

  File "/home/chrislorier/envs/netconf/lib/python3.5/site-packages/paramiko/kex_group1.py", line 111, in _parse_kexdh_reply

    self.transport._verify_key(host_key, sig)

  File "/home/chrislorier/envs/netconf/lib/python3.5/site-packages/paramiko/transport.py", line 1628, in _verify_key

    key = self._key_info[self.host_key_type](Message(host_key))

  File "/home/chrislorier/envs/netconf/lib/python3.5/site-packages/paramiko/rsakey.py", line 59, in __init__

    ).public_key(default_backend())

  File "/home/chrislorier/envs/netconf/lib/python3.5/site-packages/cryptography/hazmat/backends/__init__.py", line 74, in default_backend

    _default_backend = MultiBackend(_available_backends())

  File "/home/chrislorier/envs/netconf/lib/python3.5/site-packages/cryptography/hazmat/backends/__init__.py", line 30, in _available_backends

    for ep in pkg_resources.iter_entry_points(

AttributeError: module 'pkg_resources' has no attribute 'iter_entry_points'

I think the issue is the cryptography library is trying to set a global variable in a non thread safe way. I have found that I can get around this by calling Device.open() once before trying to parallelise, but I am not particularly confident in that approach. I also could use processes rather than threads for each device, but I would prefer not to.

How do you recommend I get around this issue?

For completeness here is the list of versions of libraries I am running:

asn1crypto                0.22.0                    <pip>

cffi                      1.9.1                     <pip>

cryptography              1.8.1                     <pip>

idna                      2.5                       <pip>

Jinja2                    2.9.5                     <pip>

junos-eznc                2.0.1                     <pip>

lxml                      3.7.3                     <pip>

MarkupSafe                1.0                       <pip>

ncclient                  0.5.3                     <pip>

netaddr                   0.7.19                    <pip>

openssl                   1.0.2k                        1

packaging                 16.8                      <pip>

paramiko                  2.1.2                     <pip>

pip                       9.0.1                    py35_1

pyasn1                    0.2.3                     <pip>

pycparser                 2.17                      <pip>

pyparsing                 2.2.0                     <pip>

pyserial                  3.3                       <pip>

python                    3.5.3                         1

PyYAML                    3.12                      <pip>

readline                  6.2                           2

scp                       0.10.2                    <pip>

setuptools                27.2.0                   py35_0

six                       1.10.0                    <pip>

sqlite                    3.13.0                        0

tk                        8.5.18                        0

wheel                     0.29.0                   py35_0

xz                        5.2.2                         1

zlib                      1.2.8                         3 

Thanks 

[Ansel Gaddy]

I have had great results using Deco module to make multiple simultaneous connections.

from deco import *
from jnpr.junos import Device

from jnpr.junos.exception import ConnectError,RpcError

@concurrent # <- this is the Deco part that makes it multithreaded

def get_disabled_lag_int_junos(**kvargs):

    dev = Device(host=kvargs['hostname'], user=kvargs['user'], password=kvargs['password'])

    xml_gotten = False

    result_attributes = False

    try:

        dev.open(gather_facts=False)

        print('{0} - connected to device {1}'.format(_get_time(),kvargs['hostname']))

        dev.timeout = 30

        xml = dev.rpc.get_configuration()

        xml_gotten = True

        dev.close()

        if 'verbose' in kvargs.keys():

            if kvargs['verbose']:

                print('{0} - got config from router {1} {2}'.format(_get_time(), kvargs['hostname'], kvargs['user']))

    except ConnectError as err:

        print('{0} - Unable to connect to router! {1} {2} {3}'.format(_get_time(), kvargs['hostname'], err, kvargs['user']))

    except RpcError as err: # attempt to retry

        print('{0} - Unable to get RPC! {1} {2} {3}'.format(_get_time(), kvargs['hostname'], err, kvargs['user']))

    if xml_gotten:

        # config section

        result_attributes_ether = identify_ether_options_inactive(xml)

        result_attributes_gigether = identify_gigether_options_inactive(xml)

        if result_attributes_ether:

            result_attributes = result_attributes_ether

        elif result_attributes_gigether:

            result_attributes = result_attributes_gigether

        # lacp section

    return {'result':result_attributes,'date':_get_time(),'hostname':kvargs['hostname']}

@synchronized

def map_get_disabled_lag_ints(**kwargs):

    results = [get_disabled_lag_int_junos(hostname=x,user=kwargs['user'],password=kwargs['password']) for x in kwargs['hosts']]

    result_gets = [x.get() for x in results]

    return(result_gets)

[Chris Lorier]

Ok, it seems like the issue here was the way pip and conda built cryptography. When I was using pip to install cryptography in my conda environment I would get:

 c/_cffi_backend.c:15:17: fatal error: ffi.h: No such file or directory

   #include <ffi.h>

Which would fail to build the wheel, so it would use setup.py instead.

To fix this I got conda to build cryptography when I created the environment rather than using pip. This seems to have fixed the issue, but conda also seems to be unable to locate cryptography 1.8.1, so I am stuck with 1.7.1 instead.

I'm not sure if the issue is with cryptography 1.8.1 or with installing cryptography with setuptools, but it seems more likely to be the latter.

[Nitin Kumar]

Hi Chris,

 

Did you install libffi-dev

https://cryptography.io/en/latest/installation/#building-cryptography-on-linux

 

Thanks

Nitin Kr

--
You received this message because you are subscribed to the Google Groups "Junos Python EZ" group.
To unsubscribe from this group and stop receiving emails from it, send an email to junos-python-...@googlegroups.com.
Visit this group at https://groups.google.com/group/junos-python-ez.
To view this discussion on the web visit https://groups.google.com/d/msgid/junos-python-ez/de65e92d-3949-4ae9-924e-1d0bcb9e0112%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.