How to enable locate and queue the database to be built

command linecronfileslocate

On a new Ubuntu 10.4 instance, I tried to use the locate command only to receive the error

locate: can not stat () `/var/lib/mlocate/mlocate.db': No such file or directory

from using this command on other systems I'm guessing that this means the database has not yet been built (it is a fresh install). I believe it is supposed to run daily, but how would I queue it up to run immediately?

Also, how is "run daily" determined? If I have a box that I only turn on for an hour at a time will the database ever be built on it's own?

Best Answer

The cron job is defined in /etc/cron.daily/mlocate.

To run it immediately:

sudo updatedb

or better

sudo ionice -c3 updatedb

This is better because updatedb is set in the Idle I/O scheduling class, so that it do not disturb (from the I/O point of view) other applications. From ionice man page:

  -c class
          The scheduling class. 0 for none, 1 for real time, 2 for 
          best-effort, 3 for idle.

  ........................

  Idle   A program running with idle io priority will only get disk time
         when no other  program  has  asked  for disk io for a defined 
         grace period. The impact of idle io processes on normal system 
         activity should be zero. This scheduling class does not take a 
         priority argument. Presently, this scheduling class is permitted 
         for an ordinary user (since kernel 2.6.25).

Related Solutions

How find only executable files using ‘locate’

Not easily. You can use

locate bash | while IFS= read -r line; do [[ -x "$line" ]] && echo $line; done

to find all executables where the name contains bash. This is faster than using find across the whole filesystem because only a few files need to be checked.

locate bash does what it always does (lists all matches)
| (pipe) takes the output from the first command (locate) and sends it to the second one (the rest of the line)
the while ...; do ... done loop iterates over every line it receives from the pipe (from locate)
read -r line reads one line of input and stores it in a variable called line (in our case, a path/file name)
[[ -x "$line" ]] tests whether the file in $line is executable
if it is, the && echo $line part prints it on your screen

Locate vs mlocate – Key Differences Explained

The locate package is the implementation of locate from GNU findutils. The mlocate package is another implementation of the same concept called mlocate. They implement the same basic functionality: quick lookup of file names based on an index that's (typically) rebuilt every night. They differ in some of their functionality beyond basic usage. In particular, GNU locate builds an index of world-readable files only (unless you run it from your account), whereas mlocate builds an index of all files but only lets the calling user see files that it could access. This makes mlocate more useful in most circumstances, but unusable in some unusual installations where it isn't run by the system administrator (because mlocate has to be setuid root), and a security risk.

Under Debian and derivatives, if you install both, locate will run the mlocate implementation, and you need to run locate.findutils to run the GNU implementation. This is managed through alternatives. If you have both installed, they'll both spend time rebuilding their respective index, but other than that they won't conflict with each other.

Best Answer

Related Solutions

How find only executable files using ‘locate’

Locate vs mlocate – Key Differences Explained

Related Question