Berkeley CSUA MOTD:Entry 42685
Berkeley CSUA MOTD
 
WIKI | FAQ | Tech FAQ
http://csua.com/feed/
2025/05/25 [General] UID:1000 Activity:popular
5/25    

2006/4/5-7 [Computer/SW/Security] UID:42685 Activity:nil
4/5     A few days ago someone asked for technical details for BART's fuckup.
        Here it is: http://www.bart.gov/news/press/news20060405.asp
        \_ well, they seem well-intentioned.  I don't think anyone's
           about to boycott Bart because of the downtime anyway.
           \_ In one of those links posted I found it interesting to read that
              one of BART's design goals is that a technical incident that
              causes the shutdown of all trains for more than 5 seconds is
              supposed to occur with a mean frequency of no more than once in
              20,000 service hours.  I think if you count the recent screwups
              as a single incident, they probably meet that goal.  Individual
              trains obviously fail more often though.
2025/05/25 [General] UID:1000 Activity:popular
5/25    

You may also be interested in these entries...
2012/5/8-6/4 [Computer/SW/Unix] UID:54383 Activity:nil
5/8     Hello everyone!  This is Josh Hawn, CSUA Tech VP for Spring 2012.
        About 2 weeks ago, someone brought to my attention that our script
        to periodically merge /etc/motd.public into /etc/motd wasn't
        running.  When I looked into it, the cron daemon was running, but
        there hadn't been any root activity in the log since April 7th.  I
        looked into it for a while, but got lost in other things I was
	...
2011/11/16-12/28 [Academia/Berkeley/CSUA, Computer/HW] UID:54230 Activity:nil
11/16   We'll be taking all CSUA machines offline in the near future for a Soda
        Hall server room reorganization (we're being moved to a neighboring
        server cabinet).  Downtime will hopefully be minimal.  --jordan
        \_ Thanks for all your work keeping the machines running!  It's
           been awesome having soda actually working again.
        \_ Update:  this is tentatively scheduled for Saturday afternoon.
	...
2011/8/9-27 [Computer/SW/Editors/Emacs, Academia/Berkeley/CSUA] UID:54162 Activity:nil
8/9     So I just found out that an old college friend of mine lost her
        four month old daughter to SIDS. What is the correct response here?
        \_ "Did she sleep with her belly?" and "Did you breastfeed?"
        \_ try this joke:
           How do you make a dead baby float?
           ˙ʎqɐq pɐǝp ÉŸo sdooÉ”s oʍʇ puɐ ɹÇ\
	...
2010/11/21-2011/1/13 [Computer/SW/Languages/Web] UID:53988 Activity:moderate
11/21   Lifehacker's recommending Dreamhost as a personal web hosting service.
        Apart from csua, who do you guys use? --erikred
        \_ What do you want to use it for? Do you need CGI or PHP?  My
           brother worked for Dreamhost and said they are unethical. In
           fact, he sued them. This refers to their treatment of customers
           and employees both. I don't know who  or what "Lifehacker" is,
	...
2010/11/1-2011/1/13 [Computer/Networking] UID:54002 Activity:nil
11/1    I'm moving from a home in Fremont to another home within the same ZIP
        code in Fremont, and AT&T customer service says I cannot transfer my
        DSL service because DSL is not available at my new home.  Is that BS?
        Are they just trying to push me to subscribe to their more expensive
        U-verse service?  I'm not asking for any lightening-speed connection.
            \_ could be
	...
2010/8/10-9/7 [Computer/Companies/Google] UID:53919 Activity:nil
8/10    http://www.businessinsider.com/google-puts-wave-out-of-its-misery-2010-8
        Google Wave No More. The people who worked on it were pretty smart.
        They wrote up a super awesome OKR with extremely low bar as a measure
        of success, exceeded everyone's expectations by going above those
        bars, and got big fat bonuses and promotions as a result of their
        planning. Brilliant.
	...
Cache (2460 bytes)
www.bart.gov/news/press/news20060405.asp -> www.bart.gov/news/press/news20060405.asp?ct=1
What follows are some details on what happened and what BART is doing to prevent similar problems in the future. BART suffered three computer shutdowns between Monday, March 27 and Wednesday, March 29. The shutdowns forced our central control staff to halt train service each time. both were related to our two-and-a-half year effort to upgrade a computer system that brings train information into BARTs control center. The first two incidents were due to a problem with the latest version of software that was installed. This software was developed, configured and thoroughly tested on a simulator before being installed for revenue service. BARTs practice for installing new software is that installation takes place when trains are not running, mostly on early weekend mornings. In fact, this software was installed over the weekend and ran without problems all day Sunday and through Monday morning peak commute. Soon after the first software problems on Monday and Tuesday, BART staff began immediately working to configure a backup system that would enable a faster recovery from any future software failure. On Wednesday, work on that backup system inadvertently contributed to the failure of a piece of hardware that, in turn, created the longest delay. It appears the technicians believed they were working in a way that would not affect service. It was a mistake to work on the computer system during revenue hours. The individuals involved did so with good intentions: to shorten the recovery time in incidents like Monday and Tuesdays shutdowns. What are we doing to correct these types of service interruptions? First, we have reverted to an earlier version of software that has provided reliable service for many months. We will continue to use this until we are confident that the new, upgraded software can perform with the same reliability. Second, we are adamant that we will not break established practice by working on the system during revenue hours. In all cases, BART holds its employees to the highest standards and will take appropriate action to ensure adherence to established procedures. Finally, we are working with the manufacturer of the computer hardware that did not perform as expected to determine what corrective steps need to be taken. BART has an excellent passenger on-time record and these service disruptions, while an anomaly, are clearly unacceptable. We sincerely appreciate your continued support and understanding.