Alerts

Posted: April 26, 2005

Due to circumstances beyond our control, the Environmental and Health Services department has informed SEAS to stop all access to Trailer F immediately. This includes card access, which has been suspended until further notice. As soon as SEAS has been told it can be re-opened, we will send out another e-mail and grant card access again. Thank you for your patience.

Posted: March 29, 2005

Due to an imminent security issue we will be turning off telnet access to the few machines that still allow telnet connections the morning of March 30th. We had planned to do that this summer anyway, this just accelerates things a bit. By now everyone should be using some form of SSH based connection to log into the CSE servers. If you still need help shifting from telnet to SSH send mail to cse-consult.

Posted: March 18, 2005

Emergency downtime tonight 3/18 @ 10pm. One of the power connectors for the system that supplies /util, /apps and /projects has failed, and must be replaced. Due to the component that failed (the connector on on of the disk chassis), the replacement requires a full shutdown of the system, and moving of the disks into a new chassis.

The unit is under a maintenance contract, and we received the replacement part today. To avoid the possibility of a second failure causing data loss, we intend to do the replacement as soon as possible. We will shut down the file server at 10pm, which will cause any operations using /projects, /apps, or /util to hang until the system comes back up. We expect that the outage will be brief - less than a half hour.

Posted: March 14, 2005

Today, all student systems (including hadar, armstrong, yeager, pollux, fork, and gagarin) will be down until at least 9pm so that home directories can be merged.

Posted: October 31, 2003

This weekend is Monthly Maintenance for the CSE systems. The big servers will go down at around 1am Sunday 11/2/2003 and should be back up by around 8:00am.

Posted: October 1, 2003

The problem with cse-consult email has been fixed, but some data for today was lost. If you sent in a support request today, we may be contacting you for additional information.

Posted: October 1, 2003

Due to a software failure the Request Tracker system that manages cse-consult email is down. Mail sent to cse-consult is being queued but will not reach the staff until the problem is fixed. Another Alert will be posted when it is fixed.

Posted: September 8, 2003

Yeager has been crashing the past few days due to a flakey problem with its memory. We think we have located which memory module was flakey and it has been removed. Hopefully that will stop the crashes.

Yeager is now running with "only" 3Gb instead of 4Gb of main memory. If this turns out to be the cause of the crashes the faulty memory module will be replaced during the next scheduled Monthly Maintenance.

Posted: August 29, 2003

Monthly Maintenance for the CSE systems will be Sunday August 31st. The large servers will go down at around 1:00am and should be back up by around 6:00am. System patches will be done to the Solaris systems so workstations in labs will be patched and rebooted as well. The patching will most likely take quite a while, lasting through until around noon of Sunday.

Please note that this is not a normal schedule for the Monthly Maintenance. The normal pattern is for Monthly Maintenance to be scheduled on the first Sunday of each month. It is being scheduled one week early this time so that it causes less of an impact on classes and to get the patches done a little earlier than next weekend.

Posted: June 13, 2003

We have begun to upgrade servers and workstations to Solaris 9. Two notable changes are that we no longer permit telnet access to CSE systems, and that an upgraded version of the Sun compilers will be installed. Announcements will be made for each server before any service outage.

Posted: May 27, 2003

University Facilities has scheduled a power outage for Bell Hall that will begin at 2am on Thursday 5/29/2003 and continue until 8am. We need to power off the computer systems during this window. The main CSE systems will begin to go down at around midnight Wednesday night. We will begin to bring the systems back up at 8am Thursday morning but it takes a long time to get the systems fully back online and functional. It would be best if you just plan to take Thursday morning off if that is an option.

At this time we are not planning to take down the workstations in Baldy or in the Trailers. But since the fileservers these systems rely on will be down they will "hang". DO NOT try to reboot them yourselves. If you just leave them alone they should recover just fine when the servers come back up. It would be best if you do not leave yourself logged in.

Posted: April 4, 2003

This weekend will be Monthly Maintenance for the CSE systems. The large central servers will go down at around 1:00am on Sunday April 6th, 2003 and should be back up by around 6:00am.

During this window I will be upgrading the Web Server machine to a faster machine with more disk space. Access to the main Web site will be flakey during this time.

Posted: February 28, 2003

This weekend is Monthly Maintenance for the CSE systems. The large servers will go down at approximately 1:00am Sunday March 2, 2003 and should be back up by around 6:00am.

Posted: February 4, 2003

There was another power failure early this morning. Almost all of the systems were back up and running by 9am but there appear to be network connectivity problems with other pieces of the campus as well as our Internet link.

Posted: February 1, 2003

A power failure hit Bell Hall around 3:00am today. Because that is the "active" time for automated system cleaning scripts and that sort of thing most of the large servers needed to have filesystem checks performed (starting around 9am when system staff were notified of the outage). Most of the critical systems were back up by 11am.

Posted: January 31, 2003

This weekend is Monthly Maintenance for the CSE systems. The big servers (armstrong, gagarin, yeager, hadar, pollux, castor, and picasso) will go down at around 1:00am Sunday February 2nd. They should be back up by around 5:00am. We will leave pegasus up and running so that long-running jobs do not die but they will "stall" while hadar is down.

During this time firmware in some of our network switches will be upgraded which requires two reboots of the switches, each reboot taking around 1 minute. All devices connected to any given switch will become unresponsive to the network during the reboots but should recover with no problems once the switch is back up.

Posted: January 27, 2003

Hadar (the grad fileserver and gateway) had a hardware panic this evening. The filesystems may be somewhat damaged because the machine did not shut down gracefully. We are in the process of bringing hadar back up and checking the filesystems. We may need to schedule an outage on Tuesday to replace the defective part.

Posted: January 7, 2003

University Facilities has scheduled a power outage for the first two floors of Bell Hall on Thursday 1/9/2003 from 4am to 8am. Since there is a chance of the power outage effecting the 3rd floor and our network connection will be cut off anyway we have decided to power down all the servers in the 3rd floor machine room as well as everything on the 1st and 2nd floors.

Beginning around 1am of Thursday 1/9/2003 we will begin shutting down the large central servers. Thursday beginning at 8am we will begin powering them back on. Please note that it takes quite a while to do this so it will be around 9am before the baseline set of services are available IF nothing goes wrong. Sometimes disk drives refuse to spin back up after power outages and that sort of thing, if there are hardware failures some services may be delayed, possibly until late morning or early afternoon.

We apologize for the inconveniences but there is nothing we can do about the needed power outage.

Posted: December 12, 2002

This weekend will be Monthly Maintenance for the CSE systems. The large servers will go down around 1:00am Sunday December 15th. We will be installing Solaris system patches on all of the Departmental machines. Even after the main servers are patched and back up the rest of the workstations should be considered "unstable", as patches get applied to them and they then reboot. Everything should be completed by around 8:00am.

Posted: November 19, 2002

As part of a renovation project that will give us a new server room in Bell, we need to take the gradlab offline starting on December 20, 2002 and lasting until January 10, 2003 (the Friday before classes start). This is necessary because the Airconditioning unit in the gradlab will be removed to be used in the new server room. The gradlab is getting new carpeting during this time as well. Card access will be disabled for the duration of this project.

Posted: October 31, 2002

This weekend is Monthly Maintenance for the CSE systems. During this weekend we will be moving /projects back to the NetApp filer from its temporary location we used when the NetApp had its hardware failure a couple of months ago.

/projects will become unavailable at around 11:00pm on Saturday November 2nd, 2002 and we *hope* will become available again mid- to late- afternoon on Sunday November 3rd. There is a LOT of data that needs to be copied so it is hard to judge exactly how long it will take.

The big servers will go down at around 1:00am Sunday November 3rd for their routine maintenance and should be back up between 4:00am and 5:00am.

Posted: October 3, 2002

This weekend will be Monthly Maintenance for the CSE systems. The large filer that failed on us about a week and a half ago has been repaired and is ready to re-enter service. I will be switching all of the machines back to using this filer for /util and /apps this weekend. We will move /projects back to this filer during the following Monthlies (November).

All of the main central servers will go down at around 1:00am on Sunday 10/6/2002. After the main servers are back up and running the workstations in the labs and offices will be rebooted to pick up the changed location of /util and /apps. Everything should be done by around 8:00am Sunday.

Posted: September 21, 2002

There was a major system failure on the server that provides /projects, /util, and /apps to almost all of the Departmental machines. It is designed so it can withstand failure of many of its pieces but so many things stopped working all at once it reached the point that one more failure would take it out of service and we would need to reload /projects from backup tapes. This would also mean any changes made to things in /projects after the last backup would be lost.

Posted: August 27, 2002

This weekend is Monthly Maintenance for the CSE systems. I will be taking the big servers (armstrong, yeager, gagarin, hadar, and pollux) down at around 1pm on Sunday 9/1/2002. They should be back up three to four hours later. I will be leaving pegasus up and running this month.

For newly arriving students... The first weekend of each month we have a "Monthly Maintenance" period. During this time we do some "health checks" on the servers, do some offline backups of some critical and/or tricky to backup filesystems (e.g. the Oracle server...), and this is when we schedule things like operating system patches, major hardware shifting, planned network outages/adjustments, etc. Since the servers go down workstations in offices and labs "hang" waiting for file transfers from the servers. PLEASE DON'T TRY TO REBOOT THE WORKSTATIONS! Once the servers come back up the workstations will continue on as if nothing happened.

Advance warning of these Monthly Maintenance periods is posted to the CSE newsgroups (sunyab.cse.general, sunyab.cse.grads, and sunyab.cse.undergrads) the week before. It is REALLY IMPORTANT that you read these newsgroups to help avoid inconveniences caused by you planning to get work done during one of these planned system outages.

If you have any questions send mail to "cse-consult@cse.buffalo.edu".

Posted: July 15, 2002

We need to move four machines in the machine room to make space for some new equipment. The four machines are alfred (main CSE Web Server), chopin, pegasus, and yeager. They will all go down around 1:00am Wednesday 7/17/2002 and should be back up about an hour later.

Posted: July 3, 2002

This weekend will be Monthly Maintenance for the CSE systems. The large fileservers (armstrong, castor, hadar, and picasso) and the large compute servers (gagarin, yeager, and pollux) will go down around 1:00am on Sunday 7/7/2002 and should be back up around 2 hours later. All other machines, including pegasus (Grad "batch" server) will remain up though they will "hang" while the large fileservers are down. Please do not try to reboot machines, just leave them alone and they will continue on after the fileservers come up as if nothing had gone wrong.

Posted: May 31, 2002

This weekend will be Monthly Maintenance for the CSE systems. The maintenance interval will begin at approximately 1:00am Sunday 6/2/2002. At that time the following machines will go down:

  file servers: armstrong, castor, hadar, picasso
  compute servers: gagarin, yeager, pollux

Posted: April 1, 2002 11:00am

Currently, network connectivity between the CSE networks in Bell hall and the rest of campus is down. CIT and CSE are working on resolving the problem. Until this is resolved, only local network activity will be possible, and incoming/outgoing email will be delayed.

Posted: Feb 18, 2002

Don't forget about the Graduate Conference Tuesday February 19th. All Graduate Students in particular are strongly encouraged to attend.

Posted: Jan 31, 2002

UB's Academic Calendar for Spring 2002 has changed from what was originally published. The new calendar is:

  Last day of classes: Friday, May 3
  NO READING DAYS
  Begin Final Exams: Saturday May 4
  End Final Exams: Thursday, May 9

Posted: Jan 31, 2002

This weekend is Monthly Maintenance for the CSE systems. All the machines will go down at around 11pm Saturday February 2nd and should be back up by around 9am Sunday February 3rd (except for picasso and the Multimedia Lab machines which will not be done until early afternoon). The maintenance interval is a bit longer than usual because we have memory upgrades that need to be installed in most of the large central fileservers.

Posted: Jan 22, 2002

At around 6am Wednesday 1/23/2002 a set of security patches that can't wait for the normal Monthlies routine will be applied to the CSE Solaris servers. After the patches complete (approximately 7am) the servers will be rebooted.

Posted: Jan 17, 2002

At around 8:00am Friday 1/18/2002 the Undergrad Lab servers (armstrong, yeager, and gagarin) will be rebooted to make some networking adjustments necessary to support the Sunray's being installed in Bell 338. Gagarin will be unstable all day Friday as it is prepared to support the Sunrays (just use yeager instead of gagarin for the day please).

Posted: Jan. 3, 2002

The weekend of Jan. 5, 2002 is a "Monthly Maintenance/Backups" weekend. All the CSE machines will go down around 11:00pm Saturday Jan 5, 2002 and should be back up around 8:00am Sunday.