News - Nov 2005

Nov 27 2004 - Tigger is down from 9AM-12Noon on Saturday Nov 27 for softwar maintenance.

Nov 19 2004 - We made a small tweak to the webmail login page. You no longer have to click on the netid box to begin entering your netid.

This change involves a tiny amount of javascript. It's possible it could adversely affect some older browsers, but the vast majority should be fine.

Nov 17 2004 - The Client Services Office, ACCC, is hiring student employees for Spring, 2005. Find all the details, including dates, deadlines and the mandatory technical test sign-up form at: http://www.uic.edu/depts/accc/cso/employment/

- Client Services Office Staff, ACCC.

Nov 12 2004 - FYI

Sorry for the late notice, but a recently ordered redundant power supply has arrived and we want to get it installed ASAP in order to prevent any repeat of the power supply failure on Wednesday. Also just arrived, and to be installed tonight, are additional memory modules that will double tigger's RAM capacity and should give a good boost to performance.

Note that we are still living with a sporadic "filesystem corruption" problem that has caused tigger to crash a few times since our upgrade to AIX V5.2. As you may know, we had applied a special IBM software patch to address this problem Wednesday morning. But that patch only made matters worse resulting in several failures Wednesday.

As Murphy would have it, the failing power supply and the bad software patch coincided, so it was difficult to determine which problem was causing the trouble. Eventually, a combination of a new power supply and removal of the bad patch led us back to where we started: still living with a sporadic file system corruption problem.

IBM is analyzing a core dump from Wednesday and hopefully they will be able to figure out why the "fix" actually made things worse, and they will fix the fix.

Thank you for your understanding and patience as we work through these issues.

Nov 10 2004 - tigger is having hardware problems. It is still up, but will continue to have intermittent problems. We have a call into our service provider for parts. We do not have an estimate at this time as to when tigger will be back up.

Nov 10 2004 - tigger was down several times Wednesday Nov. 10 between 9:15 and 16:00. The instability was caused by 2 problems: a broken power supply, and a bad software patch. The broken power supply was replaced. For several months we have been experiencing a problem which would occasionally cause corruption of a file system when the system was heavily loaded and we were waiting for IBM to resolve the problem. Recently, a patch to address this problem was made available and last night we installed it. Now it appears that the patch caused the filesystem to become unusable so we backed-out of the installation of the patch. We expect tigger to be stable again.

Nov 10 2004 - We are still experiencing intermittent problems with tigger. We do not have an estimate at this time as to when the problem will be fixed.

Nov 09 2004 - tigger will be down from 11PM until 1AM Tonight (Tuesday November 9) to apply software patches which we hope will correct the cause of the two crashes we've had in the past two days, and also correct the problem we had getting the mail spool back online after this morning's crash.

Nov 05 2004 - We found and fixed a problem that caused some email to be delayed to ECE, CS, and EVL. Some of the mail was delayed for a few days, so if you're affected, you will get the backlog all at once now.

The root cause was a networking problem that caused some, but not all, of our mail relay servers to not communicate with those particular downstream servers in ECE/CS/EVL. And since some mail was still flowing, this made it harder for people to notice they were missing mail.

Unfortunately, our normal automatic monitor that would have noticed the growing mail queues also malfunctioned, so this problem went on for far too long. We didn't have a meta-monitor, but in light of this incident, we're looking at how best to improving montioring at every level.

I believe the networking issues are resovled, and all the mail is now flowing. We apologize for this problem.

Nov 04 2004 - Blackboard was down Thursday, November 4, from around 10:30 to 14:30 and it is now working normally. We are working with the vendor to diagnose the problem but we think some contributing factors are: an unexpected restart of the database triggered accidentally by work done on another project; permissions set incorrectly on a log file but not detected until today's restart; and possible misconfiguration of the activity accumulator within the blackboard system.

Nov 01 2004 - The master machine in the argo system will be down for several hours starting at 10:00 AM on Monday, November 1. Argo account holders will not be able to login during the outage. In addition, user of ANSYS on UIC-based PCs will not be able to run the software because argo is the license server. The purpose of the downtime is to upgrade the OS on the master machine.