When a service fails, it reports what happened or when something goes wrong in your machine (for instance, an hardware issue) is also reported. Those messages can be found in the /var/log directory.  For example, many messages are reported to /var/log/syslog or /var/log/messages file.  On the other hand, if a service generates a lot of messages it’s most likely they will be written in a separated file like mail servers or Apache do.

All this is done by a daemon called syslogd, which is a system utility  providing  support for message logging on *nix systems. But logging is not reserved for system applications, you can also log your messages using syslogd, setting a proper rule in the /etc/syslog.conf file. So, let’s see how to do it…

Configuration file (/etc/syslogd.conf or /etc/rsyslogd.conf)

In the /etc/syslog.conf file (ore /etc/rsyslog.conf) are defined the rules for logging with syslogd, a rule per line. Every rule follows this syntax:

selector action

The selector part is made up of facility.priority, so the complete syntax would be:

facility.priority action

The facility is one of the following keywords:

The priority can be one of keywords listed on the following table. Al messages will be reported by priority, in ascendant order. For example, if you specify alert priority it will report alert, emer y panic priorities, but it will not report  from crit, to debug priorities.

The action field describes what to do with the reported message. Commonly all messages are written to a file know as logfile,  but there are also other actions, like forwarding messages to another host. So, the action field can be one of the following:

Modifiers

Basically there are three modifiers: =,  ! and *.  The “=” modifier makes syslogd to report only messages with the exact priority . For Instance:

mail.=error                       /var/log/mail.error

Here syslogd will report only error messages. Without the = modifier it should report error, crit, alert and panic messages. This modifier can only be used with priorities.

The second modifier is the “!”, which inverts the meaning of the rule. For example:

mail.!error                       /var/log/mail.error

Syslogd will report messages with less than error priority, ergo warning, notice, info and debug. If you want to exclude just one priority, you must use the !=  combination.

Finally, the “*” modifier lets you select among different facilities or services. For instance:

mail.*                       /var/log/mail.log

Here, all messages from mail facility will be saved to /var/log/mail.log file, not matter the priority. Another example:

*.info                       /var/log/info.log

It doesn’t matter what facility, all messages with info priority will be saved to /var/log/info.log file.

Semicolon and colon operators

The semicolon operator lets you write several rules in a more compact way. For example:

mail.=info                       /var/log/info.log
mail.=notice                     /var/log/info.log
auth.=info                       /var/log/info.log

The above rules can be written in one line:

mail.=info;mail.=notice;auth.=info   /var/log/info.log

On the other hand, if you want select several facilities, you can use the colon operator. For instance:

mail.info                       /var/log/info.log
auth.info                       /var/log/info.log

You can write the above rules in one line, like this:

mail,auth.info                  /var/log/info.log

The big difference between the semicolon and the colon operator is that the latter only separates facilities and the first one can separate facilities and priorities, even if they are not compatible.

Logging synchronously

Some logs must be monitored in real time, for instance when debugging a service. The thing is that syslogd writes messages when its buffer is full, i.e., asynchronously . If you want to write messages synchronously put a “-” before the log’s file path.

Examples

Here’s some examples I made up, and others taken from /etc/syslog.conf:

local3.info                     /var/log/mylog

Use local3 facility, reporting info messages to /var/log/mylog file.

auth,authpriv.*                 /var/log/auth.log

Report all priorities for auth and authpriv facilities to /var/log/auth.log file.

mail.warn                       -/var/log/mail.warn

Warn messages from mail will be saved into /var/log/mail.warn file synchronously.

mail.!=error                       /var/log/mail.error

All except error messages will be saved into /var/log/mail.error file.

*.=debug;\
	auth,authpriv.none;\
	news.none;mail.none	-/var/log/debug

All debug messages will be saved to /var/log/debug synchronously, except those from auth, authpriv, news and mail facilities.

The logger command

If you want to communicate to syslogd, you can use the logger command.  For instance, to send a message with info priority and local3 facility, just type:

logger -p local3.info “This is a message”

References

• man syslog
• man logger
• “Todo Linux” magazine. Year 8. Issue 92. Pages 43-47.

If you’ve never worked with tar before or if you know few options this post will show how to use tar, at least the basic options. First of all, let’s see an extract from tar’s manual description:

“Tar stores and extracts files from a tape or disk archive. The first argument should be a function; either one of the letters Acdrtux, or one of the long function names.. A function letter need not be prefixed with ‘-’, and may be combined with other single-letter options. Some options take a parameter; with the single-letter form these must be given as separate arguments. A long function name must be prefixed with –”

Ok, let’s see some of those functions and options…
Read the rest of this entry »

The netstat command

According to Wikipedia, “Netstat (network statistics) is a command-line tool that displays network connections (both incoming and outgoing), routing tables, and a number of network interface statistics”.

By default, netstat without any option shows current connection status by listing connectors. But let’s see other options you can use with the netstat command…

Options

Read the rest of this entry »

free command's output and System Monitor

Free displays the amount of free and used memory in the system, according to the manual (man free). But here’s one of the confusing terms in Linux: “used memory”. From others operating systems, we are used to the idea that “used memory” reefers to memory we already used, so we can’t use it in the future. But in Linux, “used memory” stands for memory ẗaken by the system, so it can be used by users or applications. From this point of view in Linux the bigger the used memory, the better. If you have 1 GB RAM and used memory of 900 MB, please don’t buy another extra 1 GB ram, because the more likely you will have 1.9 GB of used memory…again, don’t worry it’s normal.

Read the rest of this entry »

If you taking the Profesional Certificado Linux (PCL)  curse at ISEIT institute, here’ s a few review questions for the certification exams PCL I y PCL II (in Spanish)

• Preguntas de repaso PCL I
• Preguntas de repaso PCL II

The answers are up to you :)

Institute’s URL : http://ve.iseit.net

Tire of those non-fancy text editors? You are tied up to a specific editor? Your coworkers make laugh of you because you don’ t understand vi?…Don’t worry anymore!! Here’s a quickguide for vi.

How Vi works?

Maybe you have got tired of vi because anytime you try to use it, you don’t get the results you want. Ok, let’s talk about how vi works. It has three operation modes:

• Command mode: In this mode, vi expects an action to be performed, like copying, pasting or just moving the cursor. This is its default behavior, so when you open a file and start writing you won’t see what you are writing because you are performing commands.
• Inserting mode: In this mode you can in deed write into the file, but before you have to use one of the inserting command like “i”.
• Command extended mode: This is a special mode where you can pass arguments to commands. In order to enter to this mode just write a colon and a command. For instance:

:1,4 y

It means to copy from line 1 to 4. Now lets see some commands grouped according to their function:

Cursor movement

Read the rest of this entry »

How do you clear the screen without typing the clear command? my students always ask about this little trick I do while working on a terminal. Well, the answer is by using this shortcut: Ctrl+l. Immediately after they ask me if there are any other shortcuts, and I reply “Yes, of course…but I don’t remember them. When I start using Linux I read some of them, but I only remember the Ctrl+l shorcut”.

So, I decided it was the time to recheck those shortcuts and post them here…

Terminal’s shortcuts

This table shows some shortcuts. Keep in mind some can work o not depending on the console you use. For instance, Alt+f is a defined shortcut in Gnome’s terminal.

Shortcuts and arguments

You can pass numeric arguments to terminal commands. For example, to give the Ctrl+d command an argument of 10, you could type

, which will delete the next ten characters on the input line.

Custom shortcuts

You can go forward and define your own shortcuts in bash. To do so just add the shortcut you want on your ~/.inputrc file (or /etc/inputrc for all users). For instance, the search forward function didn’t work on my computer so I had to put this one:

“C-h”:history-search-forward

Here “C-h” means Ctrl+h. You can even define a shortcut for a command like ‘ps aux‘ adding this line:

“C-o”:’ps axun’

In order to this changes take effect  the inputrc file have to be reread using this shortcut Ctr+x Ctrl+r.

References

• http://unixhelp.ed.ac.uk/CGI/man-cgi?readline+3
• http://www.linuxquestions.org/questions/linux-general-1/any-way-to-create-bash-short-cuts-like-ctrll-for-clear-and-ctrld-for-exit-554420/

On Linux we can perform file searching in many ways. For Instance, from Nautilus (the Gnome’s file manager) we can search on the current directory for all mp3′s files using this expression ‘*.mp3′. But if we need to do more complicated searches, for example all mp3′s files greater than 5 MB, Nautilus is not the answer. In this cases we have to use a versatile and powerful command: find.

Read the rest of this entry »