US-CERT Technical Alerts and Bulletins

    Support FAQ

    print PDF
    Support FAQ
    Questions covered in this FAQ:


    General Questions

    Q. How should I submit my request for assistance?

    SSG utilizes a request tracking system (RT) to organize your requests. This system reaches all the support staff. Because a SSG person may not be in the day that you need help, it is always best to use the e-mail addresses below:

    If you are Faculty then please send technical support requests to: faculty-support@cs.ucdavis.edu

    If you are a grad student or staff then please send technical support requests to: support2@cs.ucdavis.edu

    Q. What is my password? I have forgotten my password, what do I do?

    The short answer is: "we don't know what your password is". It would take too long to find it out. What we like to do is change your password to something you do know. After we verify that you are who you say you are, and that you are the one who really owns the account you are trying to get into, we remove your old password, possibly giving you a temporary password, and ask you to enter a new one or change the temporary password.

    Q. How can I change my NIS password? How can I cange my password on the mailserver?

    You can use this command:
    yppasswd

    This will work on any system that is served by this NIS domain. To check this, visit a system where you have an account, and type this command:
    domainname

    If the system answers "fandr.nis", then using yppasswd on this system will let you update your password on the mailserver (and of course on all other hosts within the fandr.nis domain).

    If you prefer, or if you don't have any accounts within fandr.nis, you can ask us to copy a password from another system for you.

    Q. I deleted some files, what do I do?!!

    Write down the file name(s) and path(s) of the file(s) that got deleted.

    If it is a file in your home directory, go to your home directory and type

    df .

    or

    df -k .

    You should be able to tell if the drive is local to your computer, or that it is remote mounted from our home directory fileserver. If it reports back something like this line, then it is remote mounting from our Big File Server "bfs".:

    bfs6:/vol/vol1/home

    In that case, you can restore the file yourself, see the section of the FAQ on Snapshots, "How can I restore a file I removed?"

    If you don't see that it is mounting to bfs, it is probably a local drive, on your computer.

    If it is a local file, on a hard drive on your computer, send a request to support@cs asking us to attempt to restore your file. Be sure to include the name of the file, and when and from what system you deleted it.

    We perform nightly backups of most of the systems in the department. Chances are you might only lose the work you have done between the last backup and the time of the file deletion. We attempt to fulfill restore requests promptly, but we have had them take up to a week depending on emergencies, request load, and how many tapes we have to look through to find the requested files.

    Q. The system file directory is full what do I do?

    If a system directory such as /, /usr, or /tmp is full send email to the appropriate address: faculty-support@cs.ucdavis.edu or support2@cs.ucdavis.edu and let us know. If it is on a department supported system the Systems Group will fix it.

    Q. Text disappears when vi or another program sizes my xterm screen, how do I fix it?

    In Solaris you can type this command:

    eval `/usr/openwin/bin/resize`
    (note the backward facing single quotes, or "back tics")

    or add an alias to your .cshrc as so:

    alias resize 'eval `/usr/openwin/bin/resize`'
    (note there are backward and forward single quotes)

    Typing resize after that will result in the proper sizing of your COLUMNS and LINES environmental variables
    needed by vi and other applications.

    Q. Do we have a laptop that can be loaned out to me for a day or so?

    Yes. You can currently check it out by sending your email request to appropriate mail address: faculty-support@cs.ucdavis.edu or support2@cs.ucdavis.edu please include what dates you will need the laptop for.

    Q. How might I archive and compress some files in my home directory or on my machine?

    Probably the best method of archiving is to use tar and gzip (often called a tar-ball ). To "tar up" a set of files do a command like:
    tar -cvf tar_filename.tar filename1 filename2 filename3 ...

    This creates a tar archive of all the files called tar_filename.tar (you can change this name as you wish such as my_archive.tar, or my.tar, etc.)

    To archive a whole directory with tar, cd to the directory above the directory you want to archive and execute the tar command on the directory, like this:

    # pwd
    /home/mydir/Data
    # cd ..
    # tar -cvf data.tar Data
    a Data/ 0K
    a Data/projdone.dat 1K
    a Data/time.dat 1K
    a Data/projcur.dat 1K
    a Data/fas.dat 1K
    a Data/library.dat 2K
    a Data/right.dat 2K
    a Data/sgis.dat 1K
    a Data/machines.dat 2K
    a Data/vmstat. 2K
    a Data/ping-exceptions 1K
    a Data/mark_log 5K
    a Data/ping.dat 1K
    a Data/test 1K
    a Data/thirdfloor_room_inventory 2K
    a Data/printers 1K
    # ls -l data.tar
    -rw------- 1 mydir staff 29184 Feb 29 23:23 data.tar

    Note the output from tar in verbose mode (the -v option given in -cvf). If you don't need to see this, drop the v in -cvf.

    When you get your tar archive done, to compress it (or any file) just execute:

    gzip filename.tar

    gzip will compress it and add the extension .gz :

    # gzip data.tar
    # ls -l data.tar*
    -rw------- 1 mydir staff 6375 Feb 29 23:23 data.tar.gz

    To get the information out of your compressed tar archive, first uncompress it with gunzip, then un-tar it with tar (tar -xvf filename.tar):

    # gunzip data.tar.gz
    # ls -l data.tar*
    -rw------- 1 mydir staff 29184 Feb 29 23:24 data.tar
    # tar -xvf data.tar
    x Data, 0 bytes, 0 tape blocks
    x Data/projdone.dat, 333 bytes, 1 tape blocks
    x Data/time.dat, 183 bytes, 1 tape blocks
    x Data/projcur.dat, 173 bytes, 1 tape blocks
    x Data/fas.dat, 222 bytes, 1 tape blocks
    x Data/library.dat, 1174 bytes, 3 tape blocks
    x Data/right.dat, 1189 bytes, 3 tape blocks
    x Data/sgis.dat, 261 bytes, 1 tape blocks
    x Data/machines.dat, 1550 bytes, 4 tape blocks
    x Data/vmstat., 1890 bytes, 4 tape blocks
    x Data/ping-exceptions, 47 bytes, 1 tape blocks
    x Data/mark_log, 4287 bytes, 9 tape blocks
    x Data/ping.dat, 41 bytes, 1 tape blocks
    x Data/test, 223 bytes, 1 tape blocks
    x Data/thirdfloor_room_inventory, 1771 bytes, 4 tape blocks
    x Data/printers, 555 bytes, 2 tape blocks

    Note that the directory (Data)in this case is created as well, and the files put into it, because we archived the Directory. Single files put into an archive will be dropped into the current working directory.

    Please see the tar man page for more exciting details on tar!

    Q. What does CTP mean?

    Clear Text Passwords, or passwords that can be seen when someone, such as a hacker, sniffs network traffic. FTP, Telnet, rsh and other protocols use clear text passwords, so SSG is striving to remove those protocols from our network to increase network security.

    Wireless Network FAQ

    Q. How do I get a guest account for the campus wireless network (moobile net)?

    Please see: http://wireless.ucdavis.edu/guestaccess.cfm

    Q. How can I use the CS wireless network?

    Disclaimer: The setup for wireless cards is different from vendor to vendor

    First you must register the MAC address on your wireless card with the CS wireless network and dhcp server. Please send your request for wireless access in an e-mail addressed to: support2@cs.ucdavis.edu. Please include your CS account name, your full name, what room number you are primarily using the wireless device in, a phone number for that room, and the MAC address for the wireless device.

    MAC addresses are usually written on the back of a wireless card. It is a hexadecimal number that has 12 letters or numbers. It might be written like: 00034d-12a3bb or 00:01:dd:23:44:ab.

    Another way to find it might be by running the command: "winipcfg" on Windows 95/98/ME, or "ipconfig /all" from the command line on Windows NT/2000. It is listed as "Physical Address".

    Settings for wireless cards vary; here are some settings that are common to most cards:

    Our wireless network name (BSSID) is: CS wireless network

    Network type is: Access Point

    On most Windows systems, by right clicking on My Network Places and choosing properties you can access your network settings. You may have to choose a type of network, in this case wireless. Under Internet Protocol TCP/IP, properties you should choose "Obtain an IP Address Automatically". (Your machine will request network settings via DHCP from our DHCP server.)

    If you have problems:

    Check the troubleshooting guide of your wireless network card documentation.

    Check your IP settings. Make sure you have received an IP address from the DHCP server.

    You can verify this by running "winipcfg" on Windows 95/98/ME, or "ipconfig /all" from the command line on Windows NT/2000. You should have an address that is like: 169.237.6.* (the * is some number 11-249 or so.

    You can release then renew your IP address by using the commands below. Renewing your address might get you a correct address, but in some cases with Windows, you might need to reboot.

    Windows 95/98/ME:

    winipcfg /release

    winipcfg /renew

    Windows NT/2000:

    ipconfig /release

    ipconfig /renew

    Q. I'm having problems installing my wireless card on linux, do you have documents that will help?

    Try these:
    http://www.linuxdocs.org/HOWTOs/Wireless-HOWTO.html
    http://www.ibiblio.org/pub/Linux/docs/HOWTO/Wireless-HOWTO
    and, last resort, look at google:
    http://www.google.com/search?sourceid=navclient&q=wireless+linux+howto

    Q. What parts of the Engineering II building have a CS wireless signal?

    Currently we have CS wireless coverage at these locations:
    • The east wing, third floor
    • The CS administration office
    • A small amount of the north wing, second floor
    • Meeting rooms 1131 and 3085
    • Most of the research labs

    Q. What hardware and operating systems can SSG assist me with setting up wireless on?

    The CS wireless system runs on cards, so Lucent Technologies Orinoco cards will be the most compatible with the system.

    We advise the Orinoco Gold or Silver PC Card.

    Currently we can help you best with up to date Windows systems, and up to date Red Hat Linux systems. Of course we will do our best with helping you with other operating systems.

    Q. I need to load drivers for my Orinoco card on a Linux machine, do you have any tips?

    Currently there are some very good drivers for the Orinoco cards at the Orinoco site. Here are two links, the Readme file has instructions on how to load the driver:

    Linux Driver (zip)
    Readme (txt)

    The installation is fairly easy if you follow the included read me file.

    Backup and Restore

    Q. I have a local disk on my desktop or research server, how can I get it backed up?

    Please contact support with e-mail, let them know the name of the machine you want backed up and the filesystem you want backed up: support2@cs.ucdavis.edu

    Q. I have a local disk on my desktop or research server, how can I get it a restore done for it?

    Please contact support with the name of the machine and the exact filesystem directory and name of the files you wish restored: support2@cs.ucdavis.edu

    Q. I need a file restored from my NFS mounted home directory, how can I restore it myself?

    Please see the FAQ for Snapshots, under the Question: How can I restore a file I removed, or get a copy of a file I have changed, with snapshots?

    Email

    Q. What is my e-mail address?

    Your e-mail address provided by the Department of Computer Science is "youraccount@cs.ucdavis.edu", where "youraccount" represents the account name with which you login to departmental systems. (It's usually your last name, or your last name with part of your first name appended.)

    Q. How can I change where my e-mail gets delivered?

    The main mail relay system, "mailhost" (also known to mail handlers as "cs.ucdavis.edu"), was designed to contain forwarding information for your specific account. You can maintain this information by using ssh to connect to mailhost.cs.ucdavis.edu with your usual CS account name and your NIS password. Once you're in, you'll be prompted to change your forwarding information.

    Please enter a single e-mail address as your message delivery destination. Then, messages addressed to "youraccount@cs.ucdavis.edu" should be forwarded to the address you specified.

    One exception is: Users whose CS accounts reside only on the CSIF workstations cannot access the mailhost this way. Their forwarding (from the mailhost into the CSIF) is maintained automatically, and they can implement further message forwarding by creating a .forward file in their CSIF home directory.

    Q. Do we have a POP or IMAP server for faculty, staff or lecturer use?

    Yes. Faculty, staff and Lecturers can contact support2.cs.ucdavis.edu if they are interested in using it.

    Others may want to use the campus e-mail pop system, please see:

    http://itexpress.ucdavis.edu/online/

    Q. Where can I find information on how to get an e-mail list set up (Majordomo)?

    We no longer rely on aliases for supporting our e-mail lists. Instead we use Majordomo to allow the list owners to control their lists off of the mail relay. You can request a list by browsing to: http://mailhost.cs.ucdavis.edu/cgi-bin/majordomo

    (There is a Majordomo link in our Featured Links menu, to the left)

    At that link you can also modify your new lists configurations and change the list of subscribers for a list you own. If you need help click on the help button on that page for more details on how to use majordomo and the majordomo web interface.

    Q. How do I add email addresses to a Majordomo list using the web interface (MajorCool)

    Adding email addresses to your Majordomo list using the MajorCool web interface is not very intuitive, but it is easy with some quick instruction.

    First go to the MajorCool web page: http://baton.cs.ucdavis.edu/cgi-bin/majordomo

    Then click the blue “Modify” button near the top of the page.

    From the pull down list, choose your list. Then put in the password you received in the initial Majordomo list creation instructional email. Choose “List Subscribers” in the “Modify What?” section. Click the yellow GO button.

    When the next page appears it will have a box containing the current list of subscribers to this list. If this is a new list the box may be empty.

    In the box provided add, delete or modify the email addresses for this list. There should be one address per line. You should fully qualify any address not in the @cs.ucdavis.edu domain. Click the Yellow “Apply” button when you are done adding all of your addresses.

    Example:


    gribble
    dale@yahoo.com
    smith@acm.com
    jones

    This list will send email to gribble@cs.ucdavis.edu, dale@yahoo.com, smith@acm.com, and jones@cs.ucdavis.edu.

    Later, if you would like to modify this list of addresses, just follow the same instructions. When you get to the point where you first added to the list you should see the current list in the box and can modify it.

    Q. What are newsgroups, and who runs them?

    Campus runs the Usenet newsgroup services: http://email.ucdavis.edu/newsgroups/newsgroups.html

    The Usenet server, news.ucdavis.edu, maintains over 18,000 newsgroups which are organized by topic categories known as "hierarchies". If you know of a group which exists at another site but we don't seem to carry, send a request to have it added to the Usenet Administrator at usenet@ucdavis.edu.

    Q. How do I set up SSH with my mail client (such as Outlook) to send and receive SSH encrypted e-mail?

    First you will need a POP or IMAP account (you must be faculty, staff or a lecturer), please contact:
    support2@cs.ucdavis.edu

    Then you need to create an SSH tunnel between your machine and the e-mail machine. Basically this creates a private, encrypted "tunnel" between two machines. You define the ports (e-mail or web or whatever), on both machines, and then when you use those ports, the transmissions between the machines are encrypted.

    How to set up encrypted tunnels:

    The following instructions are specific to e-mail tunnels and SSH by F-Secure (Data Fellows Ltd.�). The same instructions can probably be adapted for any SSH tunneling capable software.

    Go to Edit, Properties.

    Click "Connection",
    Set "Host name or ip address:" to soda.cs.ucdavis.edu
    Set "User Name:" to your user name
    "SSH Server port number:" should be 22
    (Don't click Compression, or fiddle with SOCKS settings)

    Click "Keyboard",


    Check Auto Linefeed
    Uncheck Backspace sends Delete
    Uncheck Delete sends backspace
    Double Click "Tunneling"


    Click "Local Tunneling"
    For POP:

    Click Add

    Set as follows:
    For "Source Port", enter: 110
    For "Destination Host", enter: localhost
    For "Destination Port", enter: 110
    Click OK

    For IMAP:
    Click Add

    For "Source Port", enter: 143
    For "Destination Host", enter: localhost
    For "Destination Port", enter: 143
    Click OK
    With both IMAP and POP:
    Click Add


    For "Source Port", enter: 143
    For "Destination Host", enter: localhost
    For "Destination Port", enter: 143

    For automatic e-mail application starting:

    Click Browse, and find your e-mail application, such as:
    C:Program FilesMicrosoft OfficeOfficeOUTLOOK.EXE
    Click "OK"


    Click the next "OK"
    Save this "Session File":
    Click File, Save As
    Enter a "File name:" such as "e-mail.ssh"
    Click "Save"

    Move this to your desktop or taskbar if you like:
    Click File, Save As

    Below the "Save In:" box is a big box with your "e-mail.ssh" file. Grab (click and hold down the button) the icon next to your saved session "e-mail.ssh" file, and drag and drop it on the desktop (it will disappear from the box, don't worry about this).
    Hit Cancel
    Quit F-Secure SSH: "File, Exit"

    How to set up your e-mail client

    These instructions are specific to Microsoft's Outlook. They can easily be applied to Outlook Express, or any other e-mail tool that uses POP and/or IMAP.

    Start, Outlook, click "Tools, Accounts"
    Click "Add, Mail"
    Enter your full name, then click "Next"
    Enter your E-mail address, thus: youraccount@cs.ucdavis.edu, click "Next"
    Choose IMAP or POP (IMAP allows server side folders, POP does not)

    For "Incoming mail (POP3 or IMAP) server", enter: localhost
    For "An SMTP server� that is used for outgoing e-mail", enter: localhost

    Click "Next"
    For "Account name:" enter your CS account
    Uncheck "remember password"
    Don't check Log on using Secure Password Authentication (SPA)
    Click Next
    Choose your connection type, Phone line, if you are using a modem at home, and LAN for your connection in the office.
    Click Finish
    Quit your e-mail program
    Connecting

    Now you can connect to the e-mail server with SSH by starting SSH with the new SSH Icon on your desktop, enter your account name and your CS password to log SSH into the mail server. If you chose to automate your e-mail client starting, your e-mail application should start right after you log in, otherwise, start it now. You can minimize your SSH window, but don't exit, as your SSH tunnels are needed for your e-mail client.

    Your CS account on the mail server is strictly for e-mail; please don't use the mail server for general computing.

    Q. How can I filter out some of the Spam emails I get in the CS department?


    Spam email is notoriously difficult to avoid; however, the CS department does use software to detect it. This detection of spam email is not %100 accurate, so each user is asked to filter their email as they see fit.

    If the spam detection software is triggered by an email it will add at least two lines to the header of the email:

    X-Spam-Score: 7.647 (*******)BLANK_LINES_70_80,MISSING_MIMEOLE,NO_REAL_NAME
    X-Spam-Status: Yes

    The first line states a score (7.647) that indicates the chances that this is a spam. If this number is over a certain amount, these two lines are added. The X-Spam-Status: Yes is an indicator that this email could be spam.

    Filters can be made to filter your email using these added header lines. One tool that can do this sort of filtering is Procmail.

    Please see:
    Procmail Tips

    This is a sample Procmail recipe for moving spam to a folder into a email folder called “spam” within a UNIX directory called “mail” (in your home directory). Please follow the instructions on the Procmail Tips on testing your filters before using this example filter and procmail to filter your very important email.. Your .procmailrc file must be located on the machine that receives your email, and that machine must have procmail on it. If you need help with procmail you should contact SSG.

    #
    # Backup mail file for when testing procmail
    SHELL=/bin/sh #Use the Bourne shell (check your path!)
    MAILDIR=${HOME}/mail #Make sure this directory exists!
    # Log files are great for debugging, but they fill up quickly. If you don't want to use
    # log files, simply comment out these next two lines
    LOGFILE=${HOME}/mail/procmail.log #remember to check and delete your logs!
    LOG="--- Logging ${LOGFILE} for ${LOGNAME}, "

    #The order of the recipes is significant
    # SPAM caught in the CS department
    :0:
    * ^X-Spam-Status: Yes
    spam

    # SPAM caught by UCD campus email servers(if email is forwared from campus)
    :0:
    * ^X-UCD-Spam-Score: .*\*\*\*\*\**
    spam

    :0:
    * ^X-UCD-Spam-Score: [5-99].[0-99]*
    spam

    # Accept all the rest to your default mailbox
    :0:
    ${DEFAULT}

    Please see:
    Procmail Tips

    Printing

    Q. What are some printers are available for my use?

    There are three common printers which you can use in the department:

    • hexus Located in the basement, this printer serves the CSIF. You can only access this printer when the CSIF is open (the hours should be posted in the basement). It is automatically configured as the default printer in the CSIF. If you have problems printing with this printer, please contact: support@cs.ucdavis.edu
    • lw5 This is a HP printer located in 3085. This printer is also known as lw5duplex (for duplex printing), lw5ps and lw5plain.
    • grad This is a HP printer located in the grad access room.This printer is also known as gradduplex (for duplex printing).

    To configure your machine to print to lw5 or grad:
    remote machine that holds the lw5 queue = print.cs.ucdavis.edu
    the queues for lw5: lw5, lw5duplex, grad, gradduplex

    The lw5 and lw5duplex machines are currently a HP LaserJet 9050dn (for use when choosing drivers)

    If you need assistance setting up your printer, don't hesitate to contact:
    support2@cs.ucdavis.edu

    Q. I'm having new problems printing to a duplex printer from a Fedora 4 machine, it only prints one-sided. Is there some configuration change I need to make?

    Yes. Become root and run: system-config-printer. Double click on the printer in the Queue name list, in the Edit a print queue window, pick the Driver Options Tab. Find Double-Sided Printing and change it to: Long Edge (Standard).

    Q. How do I set up a Windows XP machine to print to an LPR printer, such as lw5?

    Please see Windows XP printing for LPR printers

    Web and Web Server FAQ

    Q. Why can't I get to my web page?

    You must set up a directory called public_html on your NFS mounted /home directory (mounted as /home in the CSIF when using the csifwww web server]. This is where the web server will map to http://www.cs.ucdavis.edu/~youraccount

    Your home directory, the public_html, and any files or directories in public_html that you want others to have access to, must be read enabled for the all (such as 755 or 644 modes for chmod).

    How do I Password Protect a Directory in my Web Space?


    Create a file named .htaccess in the web (public_html) directory you want to
    protect. So, for instance you might have a ~/public_html/Test directory you wish to password protect. You would edit the ~/public_html/Test/.htaccess file. The file content of the .htpasswd should look something like this, if your account name was gribble and you wanted the user knight and user gribble to have access:

    Deny from all
    AuthType Basic
    AuthUserFile /fandrhome/gribble/.htpasswd
    AuthName "This is a Restricted Area"
    Require user knight
    Require user gribble
    Satisfy any

    Note: For the faculty and research area on the file server, when you get to your web page from http://www.cs.ucdavis.edu/~username, the web server knows about your home directory as /fandrhome/username. If you are password protecting your web page in the csif, when you get your web page from http://wwwcsif.cs.ucdavis.edu/~username, change this line:

    AuthUserFile /fandrhome/gribble/.htpasswd

    To be something like:

    AuthUserFile /csifhome/gribble/.htpasswd

    Make sure and place the .htpasswd file in your home directory - but not in a directory served to the web (so, not in your ~/public_html directory).

    Most systems have the htpasswd command to generate encrypted passwords, if the system you are on doesn't have it, you could use the CSIF:

    Usage: htpasswd -c passwordfile username

    The -c flag creates a new file so only use it the first time, the next user name you create an entry for you should not use the -c. So, the first time use it thus:

    # htpasswd -c .htpasswd gribble
    New password: <you type in your password here>
    Re-type new password: <you type in the same password again>
    Adding password for user gribble

    Now cat your .htpasswd file and you will see something like:

    # cat .htpasswd
    gribble:y80hwpGpoPeeQ

    To add another user, such as user “knight” you would do this:

    # htpasswd .htpasswd knight
    New password:
    Re-type new password:
    Adding password for user knight

    Now it looks like:

    # cat .htpasswd
    gribble:y8/fwpGSoPPEQ
    knight:Fol7kRS0uYVmo

    Both files need to be readable by the httpd daemon:

    # cd
    # chmod 644 ~/.htpasswd
    # chmod 664 ~/public_html/Test/.htaccess

    Then you should be asked for a password if you try to get any file from the Test directory, or if you try to access the Test directory in any way.

    Some official documentation can be found at:
    http://httpd.apache.org/docs/1.3/howto/auth.html

    Q. Why is CGI turned off on the web servers here?

    It takes additional effort and ongoing education to program secure CGI scripts.

    Please see: http://www.w3.org/Security/Faq/wwwsf1.html#Q5

    Please see the following site for more details: http://www.w3.org/Security/Faq/www-security-faq.html

    Q. How can I put binary type files, such as object code, on the web for others to download?

    Most browsers allow the user to "download" or "save link as" by right clicking on the link you want to download. So, placing the files in your ~/public_html directory will allow others to download it in this way.

    A simple reminder note written on your webpage, such as: "Right click on the file of choice and select 'save link as' or 'download' to download this file.", might be a helpful addition to your index of files.

    Snapshots

    Q. What is a snapshot?


    A snapshot is a part of a WAFL file system. Our home directory fileserver uses a WAFL file system.

    A WAFL file system can "freeze frame" itself at any point in time, and make the frozen versions of the file system available via special subdirectories that appear in the current (active) file system. Each freeze frame version of the file system is called a Snapshot. This facility allows users to recover accidentally damaged or deleted data with a simple file copy operation.

    It is easy to imagine that the use of Snapshots would incur a large penalty in disk space. Each Snapshot appears to the user as though it is a read-only copy of the active file system, independently maintained on a separate area of the filer's disks. This is not how Snapshots work, and somewhat counter-intuitively, Snapshots usually require only a tiny disk space premium to maintain.

    Snapshots operate at the block level of the WAFL file system. On any type of file system, each user-visible file and directory will be comprised of some set of blocks on the physical disk media. WAFL is no different in this respect.

    Also See:
    Wikipedia's entry for WAFL
    File System Design for an NFS File Server Appliance

    Q. What can snapshots do for me in the Computer Science department?

    It's basically a backup of the home directory file systems. Both the faculty/research and CSIF home directories use this technology. In the CSIF it is done on a daily basis. The faculty and research fileserver has hourly backups. These are not done every hour, but on a special schedule designed around the workday. If you delete a file and need it back, check for it in the previous snapshot.

    Q. How can I restore a file I removed, or get a copy of a file I have changed, with snapshots?

    When you need to restore a file on one of the gbfs file system, such as the CSIF home directorys, you can simply go to the same directory as where the file is, or used to be, then type:

    cd .snapshot

    Note that the .snapshot directory is normally hidden, and cannot be seen with a generic ls command.

    That directory has a bunch of directories that are "snapshots" of the file system:

    Many hourly snapshots:
    hourly.0/ = 800 hrs
    hourly.1/ = 300 hrs
    hourly.2/ = 2300 hrs
    hourly.3/ = 2000 hrs
    hourly.4/ = 1500 hrs
    hourly.5/ = 1200 hrs
    ...

    Many daily snapshots:
    nightly.0/ = Friday
    nightly.1/ = Thursday
    nightly.2/ = Wednesday
    nightly.3/ = Tuesday
    nightly.4/ = Sunday
    nightly.5/ = Saturday
    ...

    And one or more weekly snapshot:
    weekly.0/ = On Monday\
    ...

    There is one snap shot done weekly on Sunday, and one done every day at midnight, and various hourly snapshots, on the faculty/research fileserver, are done on a pattern designed around the heaviest computer use hours in the CS dept (some done on off hours, but more done in the peak hours).

    So, It is a bit of hit and miss to find which copy of the snapshot you need if you don't know when you changed it, but if you go into those directories you can find "snapshots" of the file system for various times.

    These snapshots look just like the file system did after they were created, so you can go to your home directory and look for the file, (you may want to rename the file that is in your home directory) then use the UNIX copy command: cp, or cp -R for a directory, to copy the file or directory to where you like. You may want to use the diff command to compare your current file and the file you are about to choose from a snapshot to determine if it is what you want.

    You can use this section of the file system just like any other read only file system, so there are quite a few uses for a snapshot, but backups are it's most common use here.

    CS Cluster

    Where might I find more information about the CS Cluster?

    SSG provides two areas that contain cluster information. One area is on the cluster and is only accessable from inside the university network at: http://elvis.cs.ucdavis.edu. The other area is on this wiki in the Documents (wiki) section, called CS Cluster.

    Campus Information and Links

    Other Infomation and Links


    Further Information

    You might try our Knowledge Base page, or for further information, please see our contact page

    Meeting Maker

    The Meeting Maker server used by the CS department is at this IP address: 169.237.55.6

    System Support Group

    Created by: system last modification: Wednesday 30 of September, 2009 [19:57:45 UTC] by gribble


    News from the CSIF