Author Archive

chown symbolic link

One of the most commonly used Linux system administration tools is chown, which is part of the coreutils package. It is used to change the user and/or group ownership of a given file or directory. Something to be aware of this tool is, it doesn’t change the ownership of symbolic links, as shown below –

root@linubuvma:/tmp# touch test
root@linubuvma:/tmp# ls -l test
-rw-r--r-- 1 root root 12 Dec 20 08:01 test
root@linubuvma:/tmp# ln -s test sltest
root@linubuvma:/tmp# ls -l sltest
lrwxrwxrwx 1 root root 4 Dec 20 08:01 sltest -> test
root@linubuvma:/tmp# chown daniel:daniel sltest
root@linubuvma:/tmp# ls -l sltest
lrwxrwxrwx 1 root root 4 Dec 20 08:01 sltest -> test

The reason this doesn’t work is in the man page for chown – symbolic links named by arguments are silently left unchanged unless -h is used.” By simply running chown on symbolic link without ‘-h’ option, you are changing the ownership of the target. The ‘-h’ option affects symbolic links instead of any referenced file.

root@linubuvma:/tmp# chown -h daniel:daniel sltest

root@linubuvma:/tmp# ls -l sltest
lrwxrwxrwx 1 daniel daniel 4 Dec 20 08:01 sltest -> test

Though not portable, in some distros

 chown -R 

will recursively change the owernship of all files, including symbolic link files and directories. In my case, ‘chown -R /path/to/file’ works for GNU chown which is part of the ‘GNU coreutils 8.21’ package on Ubuntu 14.04.

Record ssh session using screen

How to record your ssh session using screen.

Per the man page – “Screen is a full-screen window manager that multiplexes a physical terminal between several processes (typically interactive shells)”.
Screen is most commonly used to create multiple sessions to remote hosts within a single terminal window or even run multiple commands locally without leaving your shell terminal. For instance, you could be tailing the log file in one session, then run a long process, then ssh into other machine etc. all within a single window.

Screen is the go to tool when setting up a remote connection, such as ssh, and you want to continue your work at any time or from any other host without worrying of a dropped connection.

In this post, I will show you how you can record your bash session.

Installation –

yum install screen        (Debian/Ubuntu)
apt-get install screen    (Redhat/CentOS)

My local environment and the remote host I am sshing to –

daniel@linubuvma:/tmp$ screen -v
Screen version 4.01.00devel (GNU) 2-May-06
daniel@linubuvma:/tmp$ uname -r
3.13.0-106-generic
daniel@linubuvma:/tmp$ cat /etc/issue
Ubuntu 14.04.5 LTS \n \l

daniel@linubuvma:/tmp$ ssh ns2 'uname -r ; cat /etc/issue'
2.6.32-642.6.1.el6.x86_64
CentOS release 6.8 (Final)
Kernel \r on an \m

The ‘-L’ option of screen is used to record your session, the session log is automatically saved in a file named ‘screenlog.n’ in your current directory.

daniel@linubuvma:/tmp$ ls
config-err-hbzs5e  one          ssh-4yheApHRgMBF  ssh-RK7GpeFuzUB8  VMwareDnD    vmware-root-2347660412
gpg-kZux7q         screenlog.0  ssh-BBblvGtb5284  vmware-daniel     vmware-root
daniel@linubuvma:/tmp$ free -m
             total       used       free     shared    buffers     cached
Mem:          3946       2489       1457          6        547       1031
-/+ buffers/cache:        911       3035
Swap:         4092          0       4092
daniel@linubuvma:/tmp$ exit
[screen is terminating]
daniel@linubuvma:/tmp$ 

The whole bash session will be logged in screenlog.0 in this case –

daniel@linubuvma:/tmp$ cat screenlog.0 
daniel@linubuvma:/tmp$ ls
config-err-hbzs5e  one          ssh-4yheApHRgMBF  ssh-RK7GpeFuzUB8  VMwareDnD    vmware-root-2347660412
gpg-kZux7q         screenlog.0  ssh-BBblvGtb5284  vmware-daniel     vmware-root
daniel@linubuvma:/tmp$ free -m
             total       used       free     shared    buffers     cached
Mem:          3946       2489       1457          6        547       1031
-/+ buffers/cache:        911       3035
Swap:         4092          0       4092
daniel@linubuvma:/tmp$ exit
exit
daniel@linubuvma:/tmp$ 

Recording your session of an ssh connection to a remote host is also similar, with ‘-L’ option followed by the command to ssh to remote host.
Option -fn (with no flow-control)
Option -t (title bar name) in this case ‘practice’.

daniel@linubuvma:/tmp$ screen -fn -t practice -L  ssh ns2
Last login: Tue Dec 27 09:46:10 2016 from linubuvma.home.net

[daniel@kauai ~]$ hostname -f
kauai.example.net
[daniel@kauai ~]$ uptime
 10:08:18 up 18 days, 10:02, 14 users,  load average: 0.19, 0.49, 0.64
[daniel@kauai ~]$ exit
[screen is terminating]


daniel@linubuvma:/tmp$ cat screenlog.0 
Last login: Tue Dec 27 09:46:10 2016 from linubuvma.home.net
[daniel@kauai ~]$ hostname -f
kauai.example.net
[daniel@kauai ~]$ uptime
 10:08:18 up 18 days, 10:02, 14 users,  load average: 0.19, 0.49, 0.64
[daniel@kauai ~]$ exit
logout
Connection to ns2 closed.
daniel@linubuvma:/tmp$ 

Additional resources –

https://www.rackaid.com/blog/linux-screen-tutorial-and-how-to/
https://linux.die.net/man/1/screen

Randomly ordering files in a directory with python

I have a playlist file which contains audio files to play. The audio player unfortunately plays the music files in a sequential order, in whatever order they are listed in the playlist file. So occasionally I have to regenerate the playlist file to randomize the audio files order. Here is a simple script that I had to write for this purpose, the core component is the random.shuffle(list) python function –

Create script file as shuffle_files.py –

#!/usr/bin/env python

import os
import random
import sys

music_files=[]

if len(sys.argv) != 2:
  print "Usage:", sys.argv[0], "/path/directory"
else:
  dir_name=sys.argv[1]
  if os.path.isdir(dir_name):
    for file_name in os.listdir(dir_name):
      music_files.append(file_name)
  else:
    print "Directory", dir_name, "does not exist"
    sys.exit(1)
# shuffle list
random.shuffle(music_files)
for item in music_files:
  print os.path.join(dir_name,item)

Run the script by providing a path to a directory with files. Each iteration should list the files in the directory in a different order.
Note – the script does not recurse into the directories, it can be easily modified with os.walk if necessary.

root@svm1010:/home/daniel/scripts# python shuffle_files.py /opt/iotop/iotop
/opt/iotop/iotop/setup.py
/opt/iotop/iotop/README
/opt/iotop/iotop/iotop
/opt/iotop/iotop/iotop.8
/opt/iotop/iotop/NEWS
/opt/iotop/iotop/iotop.py
/opt/iotop/iotop/PKG-INFO
/opt/iotop/iotop/THANKS
/opt/iotop/iotop/sbin
/opt/iotop/iotop/setup.cfg
/opt/iotop/iotop/ChangeLog
/opt/iotop/iotop/.gitignore
/opt/iotop/iotop/COPYING


root@svm1010:/home/daniel/scripts# python shuffle_files.py /opt/iotop/iotop
/opt/iotop/iotop/PKG-INFO
/opt/iotop/iotop/COPYING
/opt/iotop/iotop/iotop
/opt/iotop/iotop/setup.cfg
/opt/iotop/iotop/NEWS
/opt/iotop/iotop/README
/opt/iotop/iotop/.gitignore
/opt/iotop/iotop/setup.py
/opt/iotop/iotop/THANKS
/opt/iotop/iotop/iotop.py
/opt/iotop/iotop/ChangeLog
/opt/iotop/iotop/iotop.8
/opt/iotop/iotop/sbin


root@svm1010:/home/daniel/scripts# python shuffle_files.py /opt/iotop/iotop
/opt/iotop/iotop/THANKS
/opt/iotop/iotop/setup.py
/opt/iotop/iotop/NEWS
/opt/iotop/iotop/README
/opt/iotop/iotop/iotop.8
/opt/iotop/iotop/.gitignore
/opt/iotop/iotop/ChangeLog
/opt/iotop/iotop/sbin
/opt/iotop/iotop/PKG-INFO
/opt/iotop/iotop/iotop
/opt/iotop/iotop/COPYING
/opt/iotop/iotop/iotop.py
/opt/iotop/iotop/setup.cfg

Reference – https://docs.python.org/2/library/random.html?highlight=shuffle#random.shuffle

Ngrep is a very user friendly packet sniffer, basically the “grep” equivalent at the network layer.

Here is a quick way of figuring out the http connections your browser is making even if you are browsing to a secure site, make sure that is the only site you are visiting as the command will capture all port 80 connections.

Installation –

apt-get install ngrep

Let us redirect all traffic ngrep captured to a file –

ngrep -d any -W byline port 80 | tee  /tmp/net_output

Now visit a secure site, say https://cnet.com, you will see nicely formated output

root@lindell:~# ngrep -d any -W byline port 80 | tee  /tmp/output
interface: any
filter: (ip or ip6) and ( port 80 )
####
T 17.31.198.19:33954 -> 72.21.91.29:80 [AP]
POST / HTTP/1.1.
Host: ocsp.digicert.com.
User-Agent: Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:50.0) Gecko/20100101 Firefox/50.0.
Accept: text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8.
Accept-Language: en-US,en;q=0.5.
Accept-Encoding: gzip, deflate.
Content-Length: 83.
Content-Type: application/ocsp-request.
Connection: keep-alive.
..

From here, you can parse the /tmp/output file.

Similarly, you can parse the output file for the type of web server your favorite sites are using. Keep the ngrep command running, and visit all your favorite sites. Note, this works for http only, as https traffic is encrypted, for https only destination IP and port are shown.

In this case, I searched for the ‘Server:’ field in the HTTP response header from the web server. Apparently, nginx seems to be most popular, it is also interesting to see that AmazonS3 storage being used for hosting static content –

root@lindell:~# awk '/Server:/ {print $2}' /tmp/output |sort | uniq -c |sort -nr
    155 nginx.
     40 Apache.
     36 Apache-Coyote/1.1.
     20 Apache/2.2.3
     14 nginx/1.8.1.
      7 AmazonS3.
      6 Akamai
      5 ECS
      5 cloudflare-nginx.
      4 Omniture
      4 ESF.
      3 sffe.
      3 nginx/1.10.2.
      2 Microsoft-IIS/7.5.
      2 gws.
      2 AkamaiGHost.
      1 WildFly/8.
      1 Varnish.
      1 openresty.
      1 NetDNA-cache/2.2.
      1 Cowboy.
      1 ATS.
      1 Apache/2.2.14

References –
http://ngrep.sourceforge.net/usage.html
https://wiki.christophchamp.com/index.php?title=Ngrep

How to locate broadband Internet service providers in your area.

The FCC keeps a database of national broadband providers and it is publicly accessible at www.broadbandmap.gov. Just enter your full address or Zip code, and it will the broadband providers in your area as well as the advertised speed. One caveat is the data was last updated on June 2014, thus you might get latest information.

I checked the database for an area which had Google Fiber for the last 9 or 10 months, and it didn’t show Google Fiber as available in that area. The database has Google Fiber Inc. as a provider listed though.

If you want to check if Google Fiber is available or coming soon to your area check https://fiber.google.com/about/.

Once nice thing about the National broadband Map is the open standards API they made available to the public. It is well documented and very easy to pull data from programmatically. The API also gives you access to Census data and demographic information.

Note – most of the queries require the FIPS state and/or county codes (Federal Information Processing Standard state code). For instance, for New York state, the FIPS code is 36. Any county within a state will have FIPS county code of state FIPS code + county FIPS code. Bronx county’s (FIPS 005) full code would be 36005, for instance.

Here is a simple python script on how to interact with the API, will use Bronx county and/or NY as an example.

Let us get the overall broadband ranking within New York state –


import requests
url='https://www.broadbandmap.gov/broadbandmap/almanac/jun2014/rankby/state/36/population/wirelineproviderequals0/county?format=json&order=asc'
r=requests.get(url).json().get('Results').get('All')
for item in r:
    print item.get('rank'), item.get('geographyName')

Output based on ranking would look like this –
1 Franklin
2 Cattaraugus
3 Allegany
4 Schoharie
5 Otsego
6 Lewis
7 Washington
8 Hamilton
9 Yates
10 Delaware
11 Steuben
12 Wyoming
13 Cayuga
14 Jefferson
15 Herkimer
16 Schuyler
17 Essex
18 Seneca
19 St. Lawrence
20 Clinton
21 Montgomery
22 Chautauqua
23 Wayne
24 Columbia
25 Greene
26 Tioga
27 Livingston
28 Tompkins
29 Rensselaer
30 Chemung
31 Genesee
32 Cortland
33 Oswego
34 Sullivan
35 Albany
36 Oneida
37 Chenango
38 Orleans
39 Fulton
40 Madison
41 Niagara
42 Ontario
43 Warren
44 Schenectady
45 Ulster
46 Erie
47 Putnam
48 Onondaga
49 Saratoga
50 Broome
51 Suffolk
52 Monroe
53 Kings
54 Queens
55 New York
56 Bronx
57 Nassau
58 Westchester
59 Richmond
60 Orange
61 Rockland
62 Dutchess

Bronx county is ranked 56 out of 62, and the data for Bronx would be –

for item in r:
    if item.get('geographyId') == '36005':
        print item
        break


{u'anyWireline': 1.0,
 u'anyWirelineError': 0.0,
 u'downloadSpeedGreaterThan3000k': 1.0,
 u'downloadSpeedGreaterThan3000kError': 0.0,
 u'geographyId': u'36005',
 u'geographyName': u'Bronx',
 u'myAreaIndicator': False,
 u'population': 1482311,
 u'providerGreaterThan3': 1.0,
 u'rank': 56,
 u'stateFips': u'36',
 u'wirelineProviderEquals0': 0.0}

There is lots more you can do with the data, feel free to dig further.

In some cases, you might want to block all users from logging in to the system or just after you login, you want to prevent everyone else from connecting to the server. During server maintenance, this could be helpful or there are use cases where only one actively logged in user has to do some work if the username is a shared account.

Solution – create the /etc/nologin file, and put the text notice as the body of the file. If a user attempts to log in to a system where this file exists, the contents of the nologin file is displayed, and the user login is terminated.

[root@kauai ~]# echo 'System is under maintenance till Dec. 24, 2PM EST.' > /etc/nologin

Now try to login to the server under non super user –

daniel@linubuvma:~$ ssh ns2
System is under maintenance till Dec. 24, 2PM EST.
Connection closed by 192.168.10.103

If your ssh configuration allows it, root user can login to the server though, the root user will still be greeted with the contents of /etc/nologin file though –

daniel@linubuvma:~$ ssh root@ns2
root@ns2's password:
System is under maintenance till Dec. 24, 2PM EST.
Last login: Sat Dec 12 01:11:35 2015 from linubuvma.home.net
[root@kauai ~]# 

Reference – https://docs.oracle.com/cd/E19683-01/806-4078/6jd6cjs3v/index.html