More commands for managing files
Using find to locate files
Find command is used to locate a files in the linux file system. The command requires two arguments:
- the name of the directory where the file is looked for
- the search condition.
The basic syntax of the command is:
find directory search_condition
The search condition is normally based to the name of the file( -name value), but you can also use options that refer to dates or access settings. The find command can also have a third argument that defines, what operation is performed to the found files. The default action, that is used if no command argument is given, is -print that prints the path and name of the matching files. Following sample command would look for file called dataset27.dat from the current directory. In this case, the file is found from a sub directory dataset3.
kkayttaj@puhti-login11:~> find ./ -name dataset27.txt
./dataset3/dataset27.txt
You can also use wildcards in the name search conditions. Note however that in such case you must quote the search condition. Following command locates from your home directory ($HOME) all files that have extension .tmp.
find $HOME/ -name "*.tmp"
In the last find command examples we use -mtime search condition, that picks files based on their modification date. With a following command you can check, what files have not been accessed in drectory /scratch/project_2001234 during the last 28 days:
find /scratch/project_2001234 -mtime +28
Here the +28 means "more than 28 days". In the same ways minus character (-) means less than. So to see what files have been modified in your current directory less than 24 hours ago, you could use command:
find ./ -mtime -1
File command tells the file type
File* command evaluates the type of the given file. The syntax of the command is:
file file_name
The command prints the name of the file and a one line description of the file type. The file command recognizes most common text file formats, compressed files and linux executables. It also studies the content of the file and tries to estimate e.g. if a normal text file contains program code or some commonly used data formatting types like XML. Note however, that file often fails to classify correctly application specific files. If the file is a binary file, that is not recognized by the file command, it is reported to be a data file.
In the example below, file types of all the files in the current working directory are listed.
kkayttaj@puhti-login11:~> file ./*
./a.out: ELF 64-bit MSB MIPS-IV executable, MIPS, version 1
./common.py: a python script text executable .
/data_old.gz: gzip compressed data, from Unix
./data.txt: ASCII text
./instrction.html: HTML document text
./molecule.msv: data
./output4.jpg: JPEG image data, JFIF standard 1.01
./outout4.png: PNG image data, 640 x 480, 4-bit colormap
./output4.xml: XML document text
./poster1.pdf: PDF document, version 1.4
./report.doc: Microsoft Office Document
Count rows and characters with wc
Command wc (Word Count) is a tool that can be used to count characters (-m), words (-w) or rows (-l) that a linux text file contains. The most common use of wc command is to quickly check the row count of your file:
wc -l file_name
Another common use is checking how many rows output of a command contains. For example following command would give the number of files with extension .dat in the current directory.
ls *.dat | wc -l
Comparing two files with diff
Diff command can be used to compare two files. Diff goes through the files row by row and prints out lines that are not identical. Diff is most useful, when you need to compare two nearly identical files like two versions of the same program file. The basic syntax of the command is:
diff file1 file2
Using checksums to verify successful data storage or transfer
Checksums provide a tool to make sure that a data file is fully conserved during storage or copying. The idea behind checksums is an algorithm that calculates a number or a string, based on the content of the file. A checksum string is calculated and stored before the file is moved to a storage media or copied to a new location. Later on, when the data is retrieved from the storage or the copying process is finished, a new checksum is computed based on the retrieved or copied files. If the new checksum equals to the previously computed one, we can be pretty sure that the data is fully conserved.
One of the most common checksum algorithms is md5 that is often used to verify the correctness of data files. For example many scientific data sets, available for download in the internet, are accompanied by a list of md5 sums. The md5 sum is always a text string 32 characters long. This string has the characteristics of a good checksum: it does not tell anything about the actual content of the source file and any modification to the original file produces a completely different checksum. Other frequently used checksum algorithms include SHA (Secure Hash Algorithm) that is often used in cryptography and CRC (Cyclic redundancy check) that is common in data transport.
The example below shows, how to use md5sum in the CSC environment.
An md5 checksum for a file is calculated with command:
md5sum file_name
For example:
kkayttaj@puhti-login11:~> md5sum poster1.pdf
cc494699398122a6b6d93a5a69bd2667 poster1.pdf
You can easily store the checksum to a file by redirecting the output of the command to a new file with > character.
md5sum poster1.pdf > poster1.pdf.md5
The command above stores the checksum and file name to a new file called poster1.pdf.md5.
Checking a set of files against an md5 sum list is done by using option -c.
md5sum -c checksum_list
For example to check validity of file poster1.pdf with the previously created checksum file poster1.pdf.md5 could be done with command:
kkayttaj@puhti-login11:~> md5sum -c poster1.pdf.md5
poster1.pdf: OK
Encrypting files with gpg
If you need to work with confidential data at CSC, you can use file encryption to increase the security of your data. In normal conditions encrypting files that locate at CSC is not needed. The files can by default be accessed only by the user him/her self. In principle, the the system administrators of CSC are able to read all data at the servers of CSC. In some occasions the administrators may need to check file names and sizes, but the administrator policy of CSC strictly prohibits accessing the contents of customers data files. However, encryption may be reasonable if you for example need to copy the data outside CSC or if encryption is required by the owner of the data.
At CSC, you can use program gpg to encrypt your files. Gpg is frequently used for creating encryption key pairs to protect emails and other data transport. However, in this chapter we demonstrate only how pgp can use used to encrypt individual files.
The basic syntax for encrypting a file with gpg is:
gpg -c file_name
The command asks the user to define a password for the file. This password is not, and should not be, in any sense related to your CSC password. After confirming the password the command makes an encrypted copy of the given file. By default the encryption is done with CAST5 algorithm, but several other algorithms can be used too.
To open a gpg encrypted file, give command.
gpg gpg_cypterd_file
Gpg example
Say we have file a patients.txt that we want to encrypt. This can be done with command:
gpg -c patients.txt
When the command is launched, following prompt appears:
Enter passphrase:
Now you can type in any password for the file. In this case we use following password: y8kIeg%a. Once the password is typed, the programs asks you to confirm the password:
Repeat passphrase:
When the encryption is finished, the we now have two files: the original file and and its' encrypted version that has an extension .gpg.
kkayttaj@puhti-login11:~> ls -l
-rw-------+ 1 kkayttaj csc 1291176 Feb 11 15:57 patients.txt
-rw-------+ 1 kkayttaj csc 313848 Feb 11 16:05 patients.txt.gpg
Note that in this case the encrypted file is smaller than the original one. Now we can remove the original file.
rm patients.txt
Later on, for example after copying the file to some other location, you can extract the data with command:
gpg patients.txt.gpg
The program now asks for the password you used in encryption ( in this case: y8kIeg%a). After this you again have two files the encrypted file patients.txt.gpg and the original, readable, file patients.txt. Note that if you forget the password of your encrypted file, there is no one who can open the file!
Managing access permissions of files and directories
Hundreds of users use the computing and storage environments of CSC. To keep the files private and in order, each file and folder in the linux environment of CSC is owned by a certain user account. In linux systems each file has three user categories: owner, group and others. For each or these user categories there is three accession settings: reading, writing and execution permissions.
By default only the owner of the file can read and modify (i.e. write) the files and directories he/she has created. Other users do not have any access permissions to the files. Normally this setting is good as it keeps your data private. However, if you wish to share some data or execute self written programs the access permissions need to be modified.
Note that the project specific disk areas in Puhti and Mahti supercomputers are an exception to this rule.̣ There by default also other project members, belonging to the same unix group, have full rights to files created by other users.
You can check the access permissions with command ls -l. Let's take a look to the sample file listing that was previously used in the [ls -la example]. In this file listing the characters from second to the tenth character include the information of the access permissions. The same information is shown in the Owner filed in the file manager tool of the Scientist's User interface.
The first three of these accession characters display the permissions of the owner, next three ones display the access permissions for the linux group members and the last three characters for all the other users. Below is a sample output for ls -la command:
total 26914
drwx------+ 3 kkayttaj csc 10 Dec 22 09:12 .
drwxr-xr-x 20 root root 0 Dec 22 09:12 ..
drwx------+ 42 kkayttaj csc 472 Dec 22 09:07 ..
-rwxr-x---+ 1 kkayttaj csc 1648 Dec 22 09:01 .cshrc
-rw-------+ 1 kkayttaj csc 93 Dec 22 09:01 .my.cnf
-rw-------+ 1 kkayttaj csc 48 Dec 22 09:05 Test.txt
-rw-------+ 1 kkayttaj csc 878849 Jan 19 2009 input.table
drwxr-xr-x+ 2 kkayttaj csc 2 Dec 22 09:11 project1
-rw-------+ 1 kkayttaj csc 26432051 Dec 22 09:08 results.out
-rw-------+ 1 kkayttaj csc 25 Mar 27 2009 sample.data
-rw-------+ 1 kkayttaj csc 49 Mar 27 2009 test.txt
In the case of file Test.txt the setting is: rw-------. This means that the owner of the file (kkayttaj) has permission to read (r) and write (w) to the file. Other users have no permissions for this file. In the case of file .cshrc the definition is: rwxr-x---. In this case the owner has also execution permissions(x) to the file and also the other users that belong to group csc have permission to read (r) and execute (x) the file.
Managing access permissions in command line
In command line usage, access permissions can be modified with command chmod. This command needs two arguments:
- a string that defines what changes are to be done and
- an argument that defines the target file or directory.
In the first argument, you first define the user category: u (user i.e. owner), g (group) or, o (others). Then you define with plus or minus character if you are going to add (+) or remove (-) permissions. Finally you define, what permissions are added or removed. For example, to allow all the group members to read file Test.txt you should give command:
chmod g+r Test.txt
You can check the effect with ls -l command:
kkayttaj@puhti-login11:~>ls -l Test.txt
-rw-r-----+ 1 kkayttaj csc 48 Dec 22 09:05 Test.txt
You can define several user categories and permissions in the same time. Command:
chmod go+rwx test.txt
would add all access permissions to all users to file test.txt. To remove the permissions you should change the + character to -.
chmod go-rwx test.txt
Note that by default, changing the permissions of a directory does not change the permissions of the files and subdirectories in the target directory. Thus command:
chmod g+w project1
would not allow other group members to modify the files in directory project1. You can use option -R to do the same permissions modification recursively i.e. to all files and subdirectories in the target directory:
chmod -R g+w project1
You can use command [groups] to check which groups you belong to. To see the members of a specific group, give command:
grep group_name /etc/group