Brandeis University Computer Science - Systems Operations

The Brandeis Computer Science Systems Office is staffed by Christopher Page (cpage at cs dot brandeis dot edu) and his student assistants, known as Gurus (guru at cs dot brandeis dot edu).


Announcements

  1. 2015-01-20 One of the two main shell servers, diadem.cs.brandeis.edu, will be down from 8:30 AM to 9:00 AM on Wednesday, January 21st, to attempt to resolve a hardware issue. Please log out ahead of time if possible!
  2. 2014-08-11 In order to install recent kernel updates, the two main shell servers (diadem and tiara) will be rebooted on Monday, August 18th, at 2 PM. Please log out ahead of time if possible!
  3. 2014-08-11 Most individual workstations (as well as all the public workstations) have now been upgraded to 64-bit Red Hat Enterprise Linux 6.
  4. 2014-04-24 In order to install recent kernel updates, the two main shell servers (diadem and tiara) will be rebooted on Monday, May 5th, at 4 PM. Please log out ahead of time if possible!
  5. 2013-12-04 After years of service, coeus and themis will be replaced over winter break with newer and faster systems; as part of this deployment, all public workstations will be updated to 64-bit Red Hat Enterprise Linux 6. Check the list of public workstations for updates! Please contact guru@cs.brandeis.edu if you have concerns about this process.
  6. 2012-08-12 The mail server upgrade has been completed. Please report any problems to guru@cs.brandeis.edu.
  7. 2012-08-09 The Computer Science Department's aging mail server will be replaced with a more modern system on Saturday, August 11, 2012, starting at 9 AM. The existing server will be unavailable during the migration. Existing mailboxes will be preserved and moved to the new system, which will be available at the same hostname and IP address once the migration is complete. We estimate the new mail server should be ready by 4 PM.
  8. 2012-08-09 The Volen Center will be without power for a brief period between 8 AM and 12 PM (estimated) on Sunday, August 12th, 2012. Computer Science department systems that are not located in the Volen 024 machine room will lose power at some point during this period. Workstations should be available again by Monday morning.
  9. 2011-08-12 The Volen Center will be without power from 8 AM to 6 PM (estimated) on Sunday, August 14th, 2011. Computer Science department systems will lose power for some or all of this period; please be aware that it is likely you will be unable to access them. Most infrastructure-related systems should be restored to functioning on Sunday evening, although some workstations may not be complete until Monday morning.
  10. 2011-06-11 The Berry Patch (the Computer Science lab in room 119) will be closing for renovations on 6/17/2011. The workstations in room 118 next door will remain available, as will the remote shell servers (coeus and themis); several other workstations will also be accessible for remote-only use. The department will be building a new area in Volen 104 with new hardware for the use of Computer Science students, planned to be open for fall term. If you need a workstation during summer term and do not already have one assigned, please stop by Volen 125 or email guru to discuss your needs.
  11. The public shell servers, coeus and themis, are scheduled to be rebooted the weekend of 4/9. coeus will be rebooted on Saturday morning at approximately 8 AM; themis will be rebooted on Sunday morning at the same time. Both systems are expected to be back up by 8:30 AM. Other public workstations will not be affected by this outage.
  12. 2010-12-27 As previously announced, there will be an NFS write outage starting tomorrow morning, Tuesday 2010-12-28, at 09:00 EDT. The outage is scheduled to be completed by Tuesday 2010-12-28 at 17:00 EDT. During this outage, all NFS shares will be available, but in read-only mode. Additionally, there may be brief outages of service as shares are switched from read-write to read-only, then again as mounts are transferred from one NFS server to the other. Please plan accordingly.
  13. 2010-09-03 The CS machine remains on generator backup, but the machine room cooling system is only partly on generator backup until early next week. As such, should inclement weather from Hurricane Earl cause campus to lose electrical service, it may be necessary to shut down some servers so as to prevent a potentially harmful build up of waste heat in the machine room. If this becomes the case, we will inform affected parties.
  14. 2010-05-27, 23:45 EDT We have been monitoring the temperature in the machine room using multiple remote sensors. These indicate that the ambient temperature has returned to our standard range (69-71 degrees fahrenheit). As such, we do not anticipate the need for any further service outages at this time. Machines currently powered off should be returned to service tomorrow morning. We apologize again for the inconvenience, and thank everyone affected for their patience and understanding. - CS Systems Operations team
  15. 2010-05-27, 20:45 EDT: In-room temperature has dropped below 75 degrees, and our maintenance provider has given us a thumbs up to resume introducing heat into the room (run servers), conditional on a later investigation to determine the location of a coolant leak and a secondary cleaning of rooftop coils in approximately one month's time. As such, the vast majority of machines powered down earlier today are now back online. We will be monitoring the temperature in the machine room throughout the night, and will power machines back down should it fail to stay within acceptable parameters. Those machines not yet powered back on will be brought back online some time tomorrow morning or afternoon. Thank you again for your patience and understanding.
  16. 2010-05-27, 18:57 EDT Emergency work by Brandeis Facilities Maintenance staff caused the machine room A/C to resume circulating cool air around 16:58 EDT, albeit possibly not at full volume. The resumption of conditioned air circulation combined with emergency cooling procedures and the emergency power-off of approximately 80% of the roughly 60 or so servers and disk arrays in the CS machine room (thereby eliminating the vast majority of heat-producing equipment) has caused the ambient temperature to drop to below 80 degrees, with ambient temperature currently in the high 70s. Staff from Brandeis' cooling maintenance vendor arrivated shortly before 18:00, and are currently evaluating the in-room and on-roof components of the Liebert AC unit. Once the in-room temperature drops into the low 70s and the maintenance vendor's staff give us the all-clear to increase load on the A/C unit, we will begin to power machines back on.
  17. 2010-05-27 The Liebert A/C unit which cools our machine room appears to have suffered some kind of fault. As of 15:45 EDT, ambient temperature was in excess of 85 degrees fahrenheit, with some devices reporting internal temperatures of over 90 degrees fahrenheit. As such, all non-essential servers have been powered down, emergency cooling measures are being implemented, and our maintenance contractor has been paged. We will update this page when more information becomes available. Thank you for your patience, and we apologize for the inconvenience.
  18. 2009-10-29 There will be an emergency NFS server reboot today at 13:15. We are aware that NFS is very slow: hopefully this will fix the problem.
    Update: The NFS server has been rebooted, and service is restored. Please let us know if you have any further issues.
  19. 2009-08-15 Reminder: NFS will be unavailable today from 11:30 to 19:30 EDT. For more information and updates during the outage see http://www.cs.brandeis.edu/nfs-status.txt
  20. 2009-08-08: Please note that there will be an outage of NFS service: On: Saturday, August 15th, 2009
    From: 11:30 (11:30 AM) EDT To: 19:30 (7:30 PM) EDT For all shares hosted on: poseidon.cs.brandeis.edu
    aka: home-{2,3}.cs.brandeis.edu aka: common-{2,3}.cs.brandeis.edu To: - move said server to a new power source - apply software updates released by our vendor(s) - perform regular maintenance work During this time you will be unable to access any files stored on NFS (including NFS home and shared directories) and served off of the server in question, and personal web pages (also hosted on NFS) will be disabled. If this outage will pose an undue hardship for you or for members of your research group, or if you have any questions not answered by this announcement, please contact CS Systems Operations by email (guru@cs.brandeis.edu), phone (781 73 62740), or in person (Volen 125).

    Thank you for your patience as we work to maintain and improve department systems for the benefit of all.

  21. 2009-07-23: Today's scheduled mail outage is now complete. All services have been restored.

    Anyone experiencing problems with mail or any other service should contact guru to report the problem. (note: Please do not use email to report that you suspect a delay in email delivery. If you are correct in your suspicions, such reports would obviously be delayed in delivery.)

  22. 2009-07-22: There will be a service outage to perform scheduled maintenance on the CS Dept. mail server ({imap,mail,mailhost,pop,smtp}.cs.brandeis.edu) tomorrow (July 23, 2009) from 17:30-18:30 (5:30-6:30 PM) EDT.

    During the outage you will be unable to check for new mail, or send mail via the CS mail server. Under no circumstances should any mail be lost or otherwise not reach its intended recipient because of this outage.

    If you have any questions, comments, or concerns, please do not hesitate to contact guru@cs.brandeis.edu. If you believe that the outage will prove an undue hardship for you at its currently scheduled time, please contact guru, and we will work with you to remedy the problem.

    For information updates during the outage and after, you may check /etc/motd on all public shell servers, and guru's website (http://www.cs.brandeis.edu/~guru/).

  23. 2009-04-10: Today's scheduled mail outage is now complete. All services were restored ahead of schedule. mail.cs.brandeis.edu was brought back up (with imap, pop, etc restored) at 14:04, and mail scanning was restored at 14:46. As expected, there was a slight delay in message delivery following the restoration of mail scanning and delivery as messages which had been spooled for later processing during the virus scanner upgrade were resubmitted for scanning, but on time message delivery resumed after approximately 35 minutes.

    With all services now restored, anyone experiencing problems with mail or any other service should contact guru to report the problem. (note: Please do not use email to report that you suspect a delay in email delivery. If you are correct in your suspicions, such reports would obviously be delayed in delivery.)

  24. 2009-04-09: There will be a service outage to perform scheduled maintenance on the CS Dept. mail server ({imap,mail,mailhost,pop,smtp}.cs.brandeis.edu) tomorrow (April 10, 2009) from 14:00-15:00 (2:00-3:00 PM) EDT.

    During the first 15 minutes of outage you will be unable to check for new mail, or send mail via the CS mail server. During the last 45 minutes of the outage you will be able to read existing mail, but will be unable to send or receive new mail. Mail received (either inbound or outbound) during this time will be spooled for delivery after the outage is complete. Under no circumstances should any mail be lost or otherwise not reach its intended recipient because of this outage.

    If you have any questions, comments, or concerns, please do not hesitate to contact guru@cs.brandeis.edu. If you believe that the outage will prove an undue hardship for you at its currently scheduled time, please contact guru, and we will work with you to remedy the problem.

    For information updates during the outage and after, you may check /etc/motd on all public shell servers, and guru's website (http://www.cs.brandeis.edu/~guru/).

  25. 2009-04-04, 14:35, EDT: The department experienced an unplanned outage of NFS service today from approximately 12:07:17 EDT to 13:56:17 EDT due to a problem on our primary NFS server. During this time NFS homedirectories and personal webpages were unavailable. Service has since been restored, and we believe that most clients recovered on their own. If you experience any problems (stale file handle warnings, NFS share unavailability, login difficulties, etc), please let guru know. Apologies for the inconvenience.

  26. 2008-05-18, 03:30 AM, EDT: All Service complete, all NFS clients appear to be behaving normally. Please report any problems to guru

  27. 2008-05-12: Please note that there will be an NFS Service Outage from this Saturday, May 17 2008, 22:00 to this Sunday, May 18 2008, 06:00.

    For more information, or for updates before, during, and after the outage please check /etc/motd on public shell servers, or visit http://www.cs.brandeis.edu/nfs_status.txt

  28. 2008-05-04: Today's mail service maintenance was completed successfully and within the scheduled outage maintenance window. All mail services have been restored, and no mail should have been lost. Please feel free to contact guru with any questions, comments or concerns you may have. Thank you for your patience,

    - CS Systems Operations

  29. There will be a service outage to perform scheduled maintenance on the CS Dept. mail server ({imap,mail,mailhost,pop,smtp}.cs.brandeis.edu) this coming Sunday (May 4, 2008) from 13:00-14:00 (1:00-2:00 PM) EDT.

    During the outage you will be unable to check for new mail, or send mail via the CS mail server. Under no circumstances should any mail be lost or otherwise not reach its intended recipient because of this outage.

    If you have any questions, comments, or concerns, please do not hesitate to contact guru@cs.brandeis.edu. If you believe that the outage will prove an undue hardship for you at its currently scheduled time, please contact guru, and we will work with you to remedy the problem.

    For information updates during the outage and after, you may check /etc/motd on all public shell servers, and guru's website (http://www.cs.brandeis.edu/~guru/).

  30. We are aware of an uptick in "backscattered" spam which is being bounced back to older accounts at retired or legacy machines (which failover to @cs inboxes). Such messages are not a sign that your account has been hacked or otherwise compromised, as most of the "senders" we have seen are forged to addresses at machines that are no longer online.

    Unfortunately, there is little that we can do about such messages, as they represent a failure in the spam identification systems of the remote sites that initially received the spam, which are now bouncing it back to us, and not in our own spam identification infrastructure. (We see such messages not as spam, but as legitimate emails from remote mail servers - which they are - unfortunately they are emails which the remote servers should know better than to send.)

    However, should you receive an unmanageable number of such messages, please feel free to alert guru, both to help us in tracking the shape and scope of the phenomenon, and so that we might see about assisting you on a case-by-cases basis at this time. Thank you,

    - CS Systems Operations,
    2008-04-13

  31. 2007-10-23: Starting at 12:00 Noon (EDT) today, all mail scored >= 15 by SpamAssassin will be silently dropped. Neither the alleged sender or the final recipient will be notified of this action. A copy of all dropped messages will be kept in quarantine for a minimum of 14 days. Mail scoring >= 5 is considered to be spam.

    This decision was made on Wednesday, October 17th, in consultation with the Systems Policy Advisory Committee after 22 months of spam scoring (but no default filtering) showed SpamAssassin to be reliable and reasonably configured, with the occasional false positive never scoring > 10.

    Questions or concerns about this new policy should be directed to guru@cs.brandeis.edu. For more information on how we scan mail, and how you can implement personal filters, see the Mail Settings page.

  32. 2007-09-19: Use of 129.64.3.3 as a client-side DNS resolver is deprecated. The machine running at 129.64.3.3 was replaced yesterday with a new machine running different software which answers authoritative queries only. Attempts to make recursive ("client-side") queries against this machine/ip address will fail. You can check your client's DNS servers on Linux and OS X by looking in the file /etc/resolv.conf. To change your DNS servers on Linux, modify /etc/resolv.conf, or use your distribution's GUI toolset. To change your DNS servers on Mac OS X, modifying this listing in the text field located at: System Preferences > Network > <Your Active interface> (eg: "Built-In Ethernet" or "Airport") > TCP/IP > DNS Servers:. All managed clients should already be properly configured.
  33. Due to popular demand we have added a section to the Mail Settings page describing how to perform server-side sorting upon delivery of messages suspected (by the server) to be spam, using procmail. See the section titled "Dealing with spam using procmail," and please note that procmail is a very powerful tool that lets you permanently delete mail with great ease, so ask for help if, after reading the instructions, you are not sure how to use it.
  34. 2007-08-13: All Berry Patch RHEL workstation/shell servers have been upgraded to RHEL 5. Please mail guru to ask for assistance, report bugs, and request software or configuration changes.
  35. Any bugs on the patch machines, big or small, should be reported to guru at cs dot brandeis dot edu
  36. We have been working to integrate the berry patch Macintoshes into the department authentication framework. At this time, the assimilation is complete, and all patch machines should be identically configured. To use these machines, log in with your CS username and password at the gui or via ssh. The Macs mount home directories and /common, and have pine installed for reading email, XCode Tools (gcc, java tools, etc) for compiling your code, and most other common Unix utilities you should be accustomed to from linux (ssh, vim, emacs, and so on). If you experience any problems with these systems, or would like to request the install of any software not already present (for academic use only) please email guru.
  37. One might have assumed this didn't need to be said but: We will take disciplinary action against any students caught hacking our machines.
  38. If you are having trouble with java, you probably need to look at the Java notes
  39. Have an account? Need a computer? Try the List of Public Workstations
  40. Please note that there have been changes in mail settings. If your account was created *before* October 2004, and you have not already done so, you must perform the changes required in mail settings to get your email to work. All accounts created after October 2004 are already properly configured.
  41. You must ssh into password.cs.brandeis.edu in order to change your password.
  42. Interested in version control and change tracking for your projects? Try setting up a SVN Repository in your own homedirectory. Also check out the SVN Book for other helpful documentation (or just use good old 'man' and 'info'). There is also an older how-to if you wish to use CVS.

Here are our previously compiled FAQS. This is a good place to start if you have any questions. For information on how to use the printers in the patch go here. All users are encouraged to consult the Usage Policy if they have questions about suitable usage of Department resources.

You may use SSH (Secure SHell) to log into the CS systems remotely. Telnet has long been disabled. We suggest OpenSSH, which is a part of the base install of most Unix-like operating systems, including all standard linux distributions, and Mac OS X. ('man ssh' at a shell prompt for usage instructions) Alternatives include PuTTY or the SSH extension for TeraTerm for Windows, and MacSSH for Mac OS Classic systems. Consult OpenSSH.com for a more comprehensive list of other options and alternatives.

If you are looking for help using a unix-like operating system, which is what all of the machines in the cluster are running, then take a look at our Unix help pages.

Things to keep in mind:


Feel free to ask the gurus about things pertaining to the CoSci systems. To contact guru, you can:

For the Fall of 2011, the gurus are:

  • Matt Gabrenya (mattgab)
  • Emmanuel Awa (eawa)
  • You? We're hiring - contact the Systems Administrator to apply

Please do not contact the gurus directly at their personal email addresses or phone, as all such requests will be forwarded on to the official guru email account and voicemail, and doing so will merely increase the amount of time until someone can get back to you.


Valid XHTML 1.0 Strict