curl – use variables to show response times and other parameters

curl is a tool to interact with a server for transferring data. Although it supports various protocols, it is most commonly used with HTTP/S. It is sort of a browser for CLI folks and a go to tool when writing scripts to interact with servers.

In addition to transferring data, how do we show request and response parameters with curl. The answer is using variables, the complete list of variables can be found here.

Example – use “time_total” to show the total time, in seconds, that the full operation lasted.

$ curl  -w %{time_total} https://www.gcplinux.com
1.149143

It is best to add the variables in a file and use curl to reference the file for better formatting. Here I have added several http request and response variables I am interested in, such as num_connects, size_download, size_header, time_namelookup, time_pretransfer etc.

daniel@hidmo:/tmp$ cat ccurl.txt 
      url_effective:  %{url_effective}\n
       content_type:  %{content_type}\n
          http_code:  %{http_code}\n
       http_version:  %{http_version}\n
       num_connects:  %{num_connects}\n
      num_redirects:  %{num_redirects}\n
          remote_ip:  %{remote_ip}\n
      size_download:  %{size_download}\n
        size_header:  %{size_header}\n
    time_namelookup:  %{time_namelookup}\n
       time_connect:  %{time_connect}\n
    time_appconnect:  %{time_appconnect}\n
   time_pretransfer:  %{time_pretransfer}\n
      time_redirect:  %{time_redirect}\n
 time_starttransfer:  %{time_starttransfer}\n
                    ----------\n
         time_total:  %{time_total}\n


daniel@hidmo:/tmp$ curl -H 'Cache-Control: no-cache' -L -w "@ccurl.txt" -o /dev/null -s https://www.gcplinux.com
      url_effective:  https://gcplinux.com/
       content_type:  text/html; charset=UTF-8
          http_code:  200
       http_version:  1.1
       num_connects:  2
      num_redirects:  1
          remote_ip:  162.247.79.246
      size_download:  71273
        size_header:  537
    time_namelookup:  0.008585
       time_connect:  0.082511
    time_appconnect:  0.264110
   time_pretransfer:  0.264293
      time_redirect:  1.287257
 time_starttransfer:  3.077526
                    ----------
         time_total:  3.177939

As far as time related parameters, listed below are the ones you will most likely use –

• time_appconnect The time, in seconds, it took from the start until the SSL/SSH/etc connect/handshake to the remote host was completed. (Added in 7.19.0)
• time_connect The time, in seconds, it took from the start until the TCP connect to the remote host (or proxy) was completed.
• time_namelookup The time, in seconds, it took from the start until the name resolving was completed.
• time_pretransfer The time, in seconds, it took from the start until the file transfer was just about to begin. This includes all pre-transfer commands and negotiations that are specific to the particular protocol(s) involved.
• time_redirect The time, in seconds, it took for all redirection steps including name lookup, connect, pretransfer and transfer before the final transaction was started. time_redirect shows the complete execution time for multiple redirections. (Added in 7.12.3)
• time_starttransfer The time, in seconds, it took from the start until the first byte was just about to be transferred. This includes time_pretransfer and also the time the server needed to calculate the result.
• time_total The total time, in seconds, that the full operation lasted.

References –

https://curl.haxx.se/docs/manpage.html

https://stackoverflow.com/questions/18215389/how-do-i-measure-request-and-response-times-at-once-using-curl

How to zip or compress a folder or directory in Linux

In Linux or similar Operating Systems, zip utility is used to package and compress (archive) files.

Let us get straight to action, we have a folder to compress with zip tool –

daniel@hidmo:/tmp/tutorial$ tree .
.
??? zip-tutorial
    ??? chapter-1
    ?   ??? content
    ??? chapter-2
    ?   ??? readme
    ??? zip.txt

daniel@hidmo:/tmp/tutorial$ zip -r tutorial.zip zip-tutorial/
  adding: zip-tutorial/ (stored 0%)
  adding: zip-tutorial/zip.txt (deflated 55%)
  adding: zip-tutorial/chapter-2/ (stored 0%)
  adding: zip-tutorial/chapter-2/readme (deflated 55%)
  adding: zip-tutorial/chapter-1/ (stored 0%)
  adding: zip-tutorial/chapter-1/content (deflated 57%)

Basically we use “zip -r DESTINATION-FILE.ZIP FOLDER-TO-COMPRESS” to compress directory. Or in short “zip -r DESTINATION-FILE DIRECTORY-TO-COMPRESS“, we can skip the .zip extension.

daniel@hidmo:/tmp/tutorial$ zip -r tutorial zip-tutorial/
updating: zip-tutorial/ (stored 0%)
  adding: zip-tutorial/zip.txt (deflated 55%)
  adding: zip-tutorial/chapter-2/ (stored 0%)
  adding: zip-tutorial/chapter-2/readme (deflated 55%)
  adding: zip-tutorial/chapter-1/ (stored 0%)
  adding: zip-tutorial/chapter-1/content (deflated 57%)

To view the contents of the compressed folder without uncompressing it –

daniel@hidmo:/tmp/tutorial$ unzip -l tutorial.zip 
Archive:  tutorial.zip
  Length      Date    Time    Name
---------  ---------- -----   ----
        0  2019-10-07 21:45   zip-tutorial/
     1202  2019-10-07 21:45   zip-tutorial/zip.txt
        0  2019-10-07 21:45   zip-tutorial/chapter-2/
     1202  2019-10-07 21:45   zip-tutorial/chapter-2/readme
        0  2019-10-07 21:44   zip-tutorial/chapter-1/
      722  2019-10-07 21:44   zip-tutorial/chapter-1/content
---------                     -------
     3126                     6 files

References –

https://linux.die.net/man/1/zip

https://superuser.com/questions/216617/view-list-of-files-in-zip-archive-on-linux

Linux – how to avoid running an alias command in shell

In some cases, you might have multiple binaries, scripts or aliases with the same name in your system. Under certain circumstances you want to run only a built-in shell command, but no an alias of the command. Here are some ways to do it.

The “ls” command is usually aliased to color the output, for instance –

$ type ls
ls is aliased to `ls --color=auto'

Precede the command with “command” or “\”

$ command ls /tmp/tutorial/
chapter-one  readme

$ \ls /tmp/tutorial/
chapter-one  readme

References –

https://www.tldp.org/LDP/abs/html/aliases.html

https://www.gnu.org/software/bash/manual/html_node/Bash-Builtins.html

Linux – Cannot assign requested address

While running a performance test on a local web service, I encountered below error –

$ ab -n 600000 -c 10000 http://localhost:8080/test
...
Benchmarking localhost (be patient)

Test aborted after 10 failures

apr_socket_connect(): Cannot assign requested address (99)

Clearly the number of concurrent requests(-n) and concurrent connections(-c) is high. But would it be possible to tweak my system so that it can handle this? Apparently yes. Doing some reading no Ephemeral port range. For a typical TCP connection, a 4-tuple of source IP/port and destination IP/port is required. In our case, the source and destination IP is fixed (127.0.0.1) as well as the destination port (8080). How many source port range do we have?

$ cat /proc/sys/net/ipv4/ip_local_port_range 
32768	60999

$ echo $((60999-32768))
28231

By increasing this port range, the system will accept more concurrent connections. Run below command under root –

root@lindell:~# echo "16000 65535" > /proc/sys/net/ipv4/ip_local_port_range
root@lindell:~# cat /proc/sys/net/ipv4/ip_local_port_range
16000	65535

The performance test now runs successfully –

$ ab -n 600000 -c 10000 http://localhost:8080/test
This is ApacheBench, Version 2.3 <$Revision: 1706008 $>
Copyright 1996 Adam Twiss, Zeus Technology Ltd, http://www.zeustech.net/
Licensed to The Apache Software Foundation, http://www.apache.org/

Benchmarking localhost (be patient)
Completed 60000 requests
Completed 120000 requests
Completed 180000 requests
Completed 240000 requests
Completed 300000 requests
Completed 360000 requests
Completed 420000 requests
Completed 480000 requests
Completed 540000 requests
Completed 600000 requests
Finished 600000 requests


Server Software:        
Server Hostname:        localhost
Server Port:            8080

Document Path:          /test
Document Length:        13 bytes

Concurrency Level:      10000
Time taken for tests:   122.307 seconds
Complete requests:      600000
Failed requests:        0
Total transferred:      78000000 bytes
HTML transferred:       7800000 bytes
Requests per second:    4905.69 [#/sec] (mean)
Time per request:       2038.449 [ms] (mean)
Time per request:       0.204 [ms] (mean, across all concurrent requests)
Transfer rate:          622.79 [Kbytes/sec] received

Connection Times (ms)
              min  mean[+/-sd] median   max
Connect:      308  848 180.0    833    3955
Processing:   293 1175 198.5   1190    1967
Waiting:       88  882 210.3    946    1738
Total:        932 2023 208.9   2018    5146

Percentage of the requests served within a certain time (ms)
  50%   2018
  66%   2085
  75%   2115
  80%   2138
  90%   2216
  95%   2298
  98%   2411
  99%   2961
 100%   5146 (longest request)


$ netstat -talpn |grep '127.0.0.1:8080' |wc -l
34241

References –

https://www.ncftp.com/ncftpd/doc/misc/ephemeral_ports.html

https://httpd.apache.org/docs/2.4/programs/ab.html

In Linux, the find command is most commonly used to search files using different criteria such as file name, size and modified time. Did you know that you can search files using inode number as well? Here is how to do it?

With “ls” we can find the inode number –

$ ls -li /etc/hosts
1576843 -rw-r--r-- 1 root root 311 Jan 21  2017 /etc/hosts

Using “-inum” option of find command, we can locate the filename and its path by its inode number.

$ find /etc -type f -inum 1576843 2>/dev/null 
/etc/hosts

$ cat $(find /etc -type f -inum 1576843 2>/dev/null)
127.0.0.1	localhost
127.0.1.1	ubuntu

References

http://man7.org/linux/man-pages/man7/inode.7.html

http://man7.org/linux/man-pages/man1/find.1.html

Contents of most text files change during the life of the file , and it is common to find yourself trying to search and replace certain text across multiple files. In Linux, this is a fairly easy task. Let us go through some of the commands you will need to perform this task and then finally construct a single liner to do the job.

• grep is your best friend when it comes to finding a string in a file. In this case we are looking for the string “REPLACEME” in current directory and across multiple files –

$ grep -r REPLACEME *
host.conf:# The "REPLACEME" line is only used by old versions of the C library.
host.conf:order hosts,REPLACEME,bind
hostname:REPLACEME
hosts.deny:ALL: REPLACEME

If we are interested only in the files which contains this particular text –

$ grep -lr REPLACEME *
host.conf
hostname
hosts.deny

• sed is a tool of choice for inline editing of files –

$ cat data 
This text will be replaced - REPLACEME
$ sed -i 's/REPLACEME/NEWTEXT/g' data 
$ cat data 
This text will be replaced - NEWTEXT

From here, there are multiple ways to skin the cat – we can loop through the files and do the replacement or we can let the commands do the replacement with a wildcard.

For loop style update -

$ for f in $(grep -lr REPLACEME *); do echo "*** File: ${f} ***" ; sed -i 's/REPLACEME/NEWTEXT/g' $f; done
*** File: host.conf ***
*** File: hostname ***
*** File: hosts.deny ***

$ grep -lr REPLACEME *

$ grep -lr NEWTEXT *
data
host.conf
hostname
hosts.deny

Actually the above for loop is redundant, sed can make changes across multiple files –

 sed -i 's/REPLACEME/NEWTEXT/g' *