Connect to Impala using impyla client with Kerberos auth

Question

I'm on a W8 machine, where I use Python (Anaconda distribution) to connect to Impala in our Hadoop cluster using the Impyla package. Our hadoop cluster is secured via Kerberos. I have followed the API REFERENCE how to configure the connection.

    from impala.dbapi import connect
    conn = connect( host='localhost', port=21050, auth_mechanism='GSSAPI',
               kerberos_service_name='impala')

We are using Kerberos GSSAPI with SASL

auth_mechanism='GSSAPI'

I have managed to install python-sasl library for WIN8 but still I encounter this error.

Could not start SASL: Error in sasl_client_start (-4) SASL(-4): no mechanism available: No worthy mechs found (code THRIFTTRANSPORT): TTransportException('Could not start SASL: Error in sasl_client_start (-4) SASL(-4): no mechanism available: No worthy mechs found',)

I wonder if I am still missing some dependencies.

If that `kerberos_service_name` actually means a Kerberos service principal, it should be something like "impala/_HOST@INSERT.YOUR.REALM.HERE" with the same Realm that is referenced in your **krb5.conf** file, and "_HOST" working as a joker for the actual host that you are connecting to. — Samson Scharfrichter, Jan 28 '16 at 17:43
And I strongly doubt that Impala is running on your PC, hence "localhost" is a joke. — Samson Scharfrichter, Jan 28 '16 at 17:44
If you run into a similar error from puresasl, you should [install the kerberos Python package](https://github.com/thobbs/pure-sasl/issues/20). — Brecht Machiels, Jun 07 '17 at 15:57

score 3 · Answer 1 · edited Jun 21 '17 at 16:38

3

Install the kerberos Python package, it will fix your issue.

edited Jun 21 '17 at 16:38

Lutz Prechelt

36,608
11
63
88

answered Jun 21 '17 at 16:11

user2905548

72
3

Sumit Kumar · Answer 2 · 2019-08-04T16:33:58.587

3

I ran into the same issue but i fixed it by installing the right version of required libraries.

Install below python libraries using pip:

six==1.12.0
bit_array==0.1.0
thrift==0.9.3
thrift_sasl==0.2.1
sasl==0.2.1
impyla==0.13.8

Below code is working fine with the python version 2.7 and 3.4.

import ssl
from impala.dbapi import connect
import os
os.system("kinit")
conn = connect(host='hostname.io', port=21050, use_ssl=True, database='default', user='urusername', kerberos_service_name='impala', auth_mechanism = 'GSSAPI')
cur = conn.cursor()
cur.execute('SHOW DATABASES;')
result=cur.fetchall()
for data in result:
    print (data)

edited Aug 04 '19 at 16:33

answered Aug 03 '19 at 18:39

Sumit Kumar

39
3

thanks, after lot of troubleshooting finally this resolved my issue. – Amit Sep 20 '19 at 20:26
@SumitKumar How did you find the right versions of the libraries? We have the same issue, just with Python 3.9, but can't seem to find the correct versions to make it work. Thanks. – Pelle Ravn Nov 08 '21 at 07:52

score 2 · Answer 3 · edited Mar 19 '20 at 15:21

2

For me, installing this package fixed it: libsasl2-modules-gssapi-mit

edited Mar 19 '20 at 15:21

Nicolas Gervais

33,817
13
115
143

answered Aug 17 '17 at 17:58

ephes

1,451
1
13
19

s_mj · Answer 4 · 2019-11-27T14:19:20.560

Try this to get tables for kerberized cluster. In my case CDH-5.14.2-1.

Make sure you have a valid ticket before running this code.

with python 2.7 having below packages.

thrift-0.9.3
thriftpy-0.3.8
thrift_sasl-0.3.0
impyla==0.14.2.2

Working Code

from impala.dbapi import connect
from impala.util import as_pandas

# 21000 is impala daemon port.
conn = connect(host='yourHost', port=21050, auth_mechanism='GSSAPI') 

cursor = conn.cursor()
cursor.execute("SHOW TABLES")
# After running .execute(), Impala will store the result sets on the server
# until it is fetched. Use the method .fetchall() to pull the entire result
# set over the network (you should only do it if you know dataset is small)
tables = cursor.fetchall()

print("Displaying list of tables")
# the result is a list of tuples
for t in tables:
    # we know that each row in SHOW TABLES result
    # should only contains one table name
    print(t[0])
    # exit() enable for only one table

print("eol >>>")

score 1 · Answer 5 · answered Oct 31 '17 at 17:11

For me, the following connection parameters worked. I did not have to install any additional packages in python.

connect(host="your_host", port=21050, auth_mechanism='GSSAPI', timeout=100000, use_ssl=False, ca_cert=None, ldap_user=None, ldap_password=None, kerberos_service_name='impala')

score 1 · Answer 6 · answered Jan 24 '19 at 14:58

To connection Impala using python you can follow below steps,

Install Coludera ODBC Driver for Impala.
Create DSN using 64-bit ODBC driver, put your server details, below is sample screen shot for same
1. Use below code snippet for connectivity
  
  import pyodbc
  
  with pyodbc.connect("DSN=impala_con", autocommit=True) as conn: ... df = pd.read_sql("", conn)

score 0 · Answer 7 · answered Jan 08 '20 at 03:01

0

python cannot connect hiveserver2

make sure you install cyrus-sasl-devel and cyrus-sasl-gssapi

answered Jan 08 '20 at 03:01

kirk wu

1
1

Connect to Impala using impyla client with Kerberos auth

7 Answers7

Linked