Bash shortcuts

Bash is the program that is executed when you connect to ssh on your server.
You can choose another shell program such as sh, ash, etc.



You can find many sites that deal with this subject:
http://www.google.fr/search?q=bash

A couple shortcuts:

• bash completes your command line by pressing "tab", for example:

• bash accepts shortcuts. if for example you frequently type ls -l you can create a shortcut and type it directly:

Now if you simply type ll, it types ls -l --color .
You can benefit and add your .bashrc file for the next time if you so desire.




System Surveillance

Once your server is connected, a verification process is done to check your server.
Once done, we have a visual result. If the visual result is not normal, this permits us to deduce very quickly



The w command allows a quick uptime: the server load average and how many people are connected to your server. We can see above that it is 3:40 am. The machine uptime is 99days. There is 1 person connected and the load average is 0.00. Everything is normal.

The load average means: how many simultaneous CPU processes are occurring.
You can read the load averages for every minute, 5 minutes, and 15 minutes.
Usually, the load average should not exceed 1 on a server with 1 CPU.
But everything depends on what is running on the system.



Freeing RAM usage. We can see that the RAM has 256 Mb of memory.
The RAM is running at almost 100%. This is normal even if nothing is running on the server.
Linux uses all available RAM in order to accelerate the execution of network storage and other diverse information processes.
When a computer process needs to use RAM, Linux will erase some the information and provide space for the process with the necessary RAM.
Now we see that 70 Mb is actually used for the process and the rest of the 184 Mb is used just for Linux.

The swap (RAM on disk) is an extra 512 Mb of RAM. This is used when there is no more RAM (actually none at all). Please consider that the swap is much slower than RAM. It is not good if your server uses swap a lot, because you can easily overload your disk and wear it out quickly. Also, the execution of processes takes longer than normal. This causes the server to slow down. And eventually the server will crash due to the ongoing requests and simultaneous processes running at the same time. It is a classic situation that faces almost all beginners. Solution? Reboot the server to hard disk. Consult the guide: How to reboot my server ?

What runs on the server ?



We can now see that the ps auxw outputs the result on several columns.
The 1st column is the user that runs the server.
The 2nd column is the pid of the process
The 3rd column is the amount of CPU usage for the process
The 4th column is the amount of RAM used for the process
The next to last column is the amount of accumulated execution time (equivalent to 100% CPU usage).
The last column is the name of the process. if you type ps auxwwww you will see the full command name.

root 1 0.0 0.1 1376 452 ? S Jun26 0:14 init [3]

init manages the full functioning of the server. It is the first executed process.
A level 6 management is permitted via /etc/inittab. The interesting/important levels are:
0 - stop or start the machine.
1 - single or repair mode (no connection)
3 - normal functioning
6 - reboot the server

It is strictly advised not to change these files. If you change them, your server may not be able to boot properly.
The network functions uniquely on level 3.

The startup files of each level can be found in /etc/rc.d/ and then for level 3 its in /etc/rc.d/rc3.d/. You can

then see what will startup (with S) and what won't (withK).



In order we find: the network startup, sshd, syslog, named, xinetd, qmail, httpd, imapd, mysql and webmin.
The other startups are administrative (crond, atd, local, etc)




the process that manages swap. If it uses too much CPU does that mean that my sever uses a lot of swap ?





These are the processes that manage the log book of your disk partitions.



We can see that your / and /home partitions are ide (hd) disks loaded in priority(hda).

/dev/shm is the disk ram that can be used. (opposite of swap).




The /dev/hda1 partition is loaded into ext3 en rw (read and write). ext3 is the log version
of the ext2 (a frequently used partition system used in Linux) The log consists of a writing an account of the information about the processes carried out on the system. In case of a hard disk reboot, the system does not need to verify the hard disk for errors. It is enough to just run the log to find any errors.

For example a df with a disk in scsi:

in raid-5 soft scsi with an adaptec card



in raid-1 hard scsi with a mylex card



in nfs (network file system)



The syslog8 process takes care of recording all the messages that the system sends. It is recorded regularly in /var/log.



With /var/log/messages you can find the messages that the principal processes send such as telnet, sshd, named etc.
With tail -f you can read a file and view newly saved data.

Who listens to what ?

Next we can view all the processes that manage the servers: apache, qmail, sshd, named
that can be recovered with netstat -tanpu




We can see, for example, that there is a process listening to port 0.0.0.0:80 that called httpd

with a pid 7798. To verify:




port 80 is the web. Is it functioning ?





We will now kill the connection:



And is it still functioning ?



It is now dead. Time to restart:




and verify some information:



>

It is operating anew.

Administration in real time

If you wish to have ps auxw permanently, you can use the top command.



Click on shift-m to sort by RAM usage. If your server crashes from using swap, launch the top command and wait until it crashes. You will see on your screen the last top command and that indicates which process made your system crash.

Voila! From now on you know what is happening on your system, how to view the processes, and how to kill them and restart
them. That's much more than just loading the machine....