This document provides information about accounts on the Unified Linux Systems managed by the School of Informatics and Computing. There are many other standalone Linux systems in the school that are not part of the unified configuration and, therefore, this document does not apply to accounts on those systems. If you have any questions about Linux systems within the school, please contact us.
The home directories on the unified linux systems are subject to disk quotas. In order to check your disk usage, we recommend you see the KB article How do I get detailed information about my Linux accounts?. This gives lots of valuable information about your account, including disk usage.
You can also try using the "quota -v" command at the linux prompt. Just keep in mind that the output may be somewhat cryptic so you are encouraged to use the above web interface. In addition, the quota command does not work at all on the workstations using NFSv4 for home directory delivery.
Effective Space Utilization
In general, we do not increase user quotas unless all other available options are exhausted and then only with the sponsorship of a member of the SoIC faculty. However, we find that most requests for more quota are not actually needed. Here are some tips for making the most of the disk space you have available.
- Utilize Existing Available Space - There are a number of storage options available to you other than your quota-enabled home directory. The options include scratch space that is like a networked /tmp, larger storage areas that are persistent but just not backed up, and fully backed-up research space. For all the details about the available options, please see the KB page What data storage space is available on the unified Linux systems?
Checkusage - You should see which directories are eating up all your quota and remove unneeded files. You can use the checkusage script to find out which files and directories are taking up the most space. If you just run checkusage with no arguments, it it will report the disk usage in your home directory. You can also name a specific directory to check with something like:
The checkusage script just uses basic unix tools like du and find which you can use manually, if you like. For example, if you cd to your home directory and run
you will get a sorted list of disk usage by directory. With this information, you can start looking for things to remove.
Core Files - Check to see if you have any "core" files laying around. A core file can be generated when a program dies a miserable death and these files tend to be very large. You can run
to list all of the core files that are in your home directory. If you are curious which program dumped core, you can run "file core.1234" to see. These core files can safely be removed.
- Web Server Cleanup - If you are running a Web Server in the burrow, there are a couple things you can do to clean up space after you install the server.
The apache src directory where you built apache can be quite large. You can run a "make clean" in that directory to clean up lots of space taken up during the build process but not needed after apache is installed. Alternatively, you can move the src directory off to your /nobackup space (see above) and leave a link behind by running something like:
- Apache creates log files that will continue to grow without bounds unless you periodically clean them up. These will be located in your apache logs directory and are typically named access_log and error_log.
- Browser Cache - Web browsers (such as firefox and opera) maintain a cache of web pages and images you visit and this cache can take up significant disk space. You can adjust the amount of disk space used for this disk cache as follows:
- Firefox - Select "Preferences" from the edit menu then select the Network tab from the Advanced category. From this page, you can adjust your cache settings and clear your disk cache. The cache files are stored in a Cache directory located in a subdirectory under ~/.mozilla/firefox.
- Opera - Select "Preferences" from the File menu and then select the "Network" category. Under the "History and Cache" tab, you can set all kinds of options for the opera cache. Opera stores the cache files in ~/.opera/cache4.
- Trash Cleanup - If you delete files using a GUI (KDE or GNOME) "Trash Can", the files may not actually be removed. You will need to purge the files from the trash can before they are actually removed. You may find these deleted files in ~/.Trash or ~/.local/share/Trash. You can also usually right click on the Trash icon and remove files.
Anaconda Cleanup - If you are using the Anaconda python version then it can leave behind lots of very large temporary files. You can clean these up by running:
Tracker .nfs files - We have see the Gnome Tracker database leave behind lots of files that eat up space. The process removes the temporary database files but leaves them open so they are not actually removed from the file server so they continue to use up quota space. The symptom is that there are a lot of large files with names that start with ".nfs" in the ~/.cache/tracker directory that can't be removed. You should be able to free up space by logging out and back in again or you can disable tracker entirely by running these two commands:
Temporary Files - You can remove unneeded temporary files. For example, emacs saves backup copies of files you edit with a ~ extension. For example, if you edit a file named foo.c, emacs will create a backup copy named foo.c~. You can find all such files by running
and can remove them by running
Compress Files - You can greatly reduce your disk usage by compressing files that are currently not being used. There are several file compression utilities available, including gzip. For example, if you have a file named somefile you can compress it to a file named somefile.gz by running "gzip somefile". You can uncompress it back into somefile with "gunzip somefile.gz". You may want to look for large files that are candidates for compression using the find command. For example, if you wanted to find all the files you have that are larger than 10,000KB you could run the following:
There is also a helpful program installed called "z" that makes file compression a breeze. For example, if you have a directory called "a343" that you are not currently using, you can just run "z -gz a343" to compress it into a file named "a343.tar.gz". If you need the directory restored to it's original state, you can just run "z a343.tar.gz". Run "z -h" for more information about using z.
One caveat to all of this. You need to have some working space available in your home directory in order to compress files since the uncompressed and compressed versions will both be in the directory during the compression process. If you are already up against your disk quota, the compression may fail. If this happens, you may want to move the file to your /scratch directory (see below), compress it there, and then move the compressed version back into your home directory.
- Strip Binaries - You can "strip" binaries to reduce their size. For example, if you have a binary called "foo" you can strip the file by simply running "strip foo". See the strip man page for more information.
We generally do not grant quota increases but if you have exhausted all other avenues and want to request a quota increase, please contact us. In your request, describe your need and we will try to help.