Due to a disk error, the RED DHCP server stopped giving out DHCP addresses to many clients and we were forced to reboot it immediately. This may also have interrupted LTSP sessions. The machine has now been recovered and both RED DHCP and LTSPs should be performing normally. We believe that the problems started sometime around 4:15pm; they were resolved by 5:50pm.
(5:50pm Thursday November 19th 2009)
Comps2, which suffered a disk failure, has been repaired and has re-entered service.
(Monday, Sept. 14 2009)
We've recently changed things so that if an unregistered machine is plugged in to our 'red' laptop network it can immediately be registered for 'red' network access; attempts to browse the web from such a machine will be redirected to a new registration system.
(Monday, August 24th 2009)
The CSLab router that forwards traffic to and from the department's various sandboxes, including the RED and LTSP networks, experienced a power failure and shutdown at about 1:15 today, Thursday, July 30th.
Since we have designed our network with such potential failures in mind, we put into service an identically configured router that was sitting on standby. This took a bit of time since we needed to cable it up and perform some final checks.
The outage lasted for about 30 minutes during which time any machine in one of the CSLab sandboxes would have been unable to access network services.
We believe that the reason the primary router failed is because one of its power supplies has failed but we are still investigating.
We apologize for the inconvenience.
(Thursday, July 30th 2009)
Please note that you should run alpine instead of pine on the Ubuntu 8.04 machines. This is a change that Ubuntu made.
There was a problem with the vacation message program which has been resolved. All mail was delivered to CSLab recipients but the sender may have received an undeliverable error message.
Initially there was an installation problem with acroread on the apps servers. This has also been resolved.
(Wednesday, July 22nd 2009)
The application (cs.toronto.edu, apps0, apps1, etc.) and computational (comps0, comps1, etc.) servers were upgraded from Ubuntu 6.06 LTS to Ubuntu 8.04 LTS.
For a period of time extending into the fall term, there will be one application server and one computational server still running Ubuntu 6.06: the application server is "oldapps.cs.toronto.edu" and the computational server is "oldcomps.cs.toronto.edu". If for some reason your software cannot run on Ubuntu 8.04, these machines will give you a place to run it for a while to give you more time to migrate it. Your point of contact (POC) will be available to help with the migration.
(Tuesday, July 21st 2009)
The network fibre from our CNS-operated gateway switch to the campus network backbone is currently unlit. CNS reports that there is a problem affecting all of Sandford Fleming, and they are looking into it.
At around 10:00am CNS had restored contact, but as of 10:10am it's gone again. We still don't have any details.
10:30am it's once again back up and should stay up according to CNS. Rogue machine(s) on the ECF network were effectively denial of service attacking shared network resources, and they have now been blocked.
(Monday, June 8th 2009)
We have just introduced a way for people to create their own VPN accounts (or to recover the password if they've forgotten it). See the self-serve VPN account web page.
(Thursday, June 4th 2009)
In response to a recent security vulnerability in our webmail software, we have upgraded the software to the latest version, which has resolved the issue. If you notice problems or changed behaviors, please let your POC know.
(Tuesday May 19th 2009 at 12:40)
CSLab is making available an apps server and a compute server that are running Ubuntu 8.04, a more recent version of the Linux distribution we use. These are called testapps2.cs and testcomps.cs respectively, and are the first steps in a planned rollout of Ubuntu 8.04 to all of our existing servers. We encourage you to try them out.
(Wednesday April 29 2009)
Our scheduled downtime to perform maintenance on our fileserver and network infrastructure is complete. Matlab 7.7 was also installed on the compute servers. See Matlab General Reference for details.
(Wednesday, Mar 4 7:30pm)
Our scheduled downtime to perform maintenance on our fileserver and network infrastructure is now complete.
(Wednesday, Feb 11 8:10pm)
We have completed migrating the fs8 filesystems back to the original fileserver. Everything should be back to normal.
(Friday, Feb 6 7:20 pm)
We experienced a double failure on one of the fileservers (fs8) today. We have brought up the fs8 filesystems on a temporary fileserver. In order to restore backup functionality the filesystems must be put back onto the original fileserver, which is in the process of being rebuilt. The downtime will begin at 5:30pm today. As well as home and work directories that are on fs8, servers and services that use fs8 may be affected.
(Friday, Feb 6 2:20 pm)
We experienced a hardware failure on one of the fileservers (fs8). Service has been restored. We will need to have a short outage to restore backup functionality.
(Friday, Feb 6 12:40 pm)
While migrating some of our network connections in preparation for the installation of a new router, we noted that a number of machines were provoking icmp redirects, which in turn caused a slowdown on some of our networks and servers. We have fixed this problem by modifying the routing tables on those machines that caused these icmp redirects and we believe everything is back to normal. Machines affected included cifs.cs, cups.cs and dcssmb.cs but other servers may have notice the occasional slowdown as well.
(Monday, Jan 26 2:00 pm)
We experienced a hardware failure with dcsweb.cs.toronto.edu. The webserver is currently back up and operational.
(Monday, Jan 26 2:30pm)
Computing and Networking Services (CNS), the central network connectivity provider for the St.George campus, will be performing an upgrade to the SF core router on Friday, January 23rd, 2009 between 7:55 and 8:05. During the upgrade the router will be down for 5 minutes. This will interrupt the Computer Science department's access the external world. For complete details on departments/networks affected please see CNS Network Events.
(Thursday, Jan 22 11:30am)
Due to the campus networking power interruption reported below, comps4.cs became unstable and we had to reboot it at 5.00 pm, January 19th. We notified all the users logged in and asked them to save their work and log out. They did so and we then reset the machine. The machine is no longer exhibiting problems.
(Monday, Jan 19 5:00pm)
Campus networking services reported a power interruption Saturday January 17 at 1:20am. This may have affected some of our networking equipment and servers, possibly causing network interruptions or reboots. We believe everything is back to normal now, but please do not hesitate to contact us should you observe otherwise.
(Monday, Jan 19 9:00am)
We have just introduced a way for people to register machines on our 'red' laptop network themselves. See the self-serve red network registration web page.
(Wednesday, January 14th 2009)