Author: Andrei Fortuna
Date: 02:45:54 07/23/04
Go up one level in this thread
Hi, I just noticed this thread. My current time schedule is very full at the moment and I can't fix the search engine bug in the next couple of months. If anyone who has spare time and knowledge on those matters is willing to help fix this problem I am willing to send him the source code of the search engine. I am a bit embarassed to make it public as it was written many years ago and I am not proud of the code style - everything was written from scratch and I didn't use STL or anything to simplify the work. I also was just beginning to put my teeth into client-server programming and as a consequence I adapted a working code for the client/server communication. The result is the mentioned fork bug. So : if anyone is kind enough to fix the bug (and not laugh at the code written by a younger instance of me) please contact me at andreif@mail.dntis.ro Best Regards, Andrei On July 22, 2004 at 10:04:09, James Swafford wrote: >On July 22, 2004 at 09:54:52, Anthony Cozzie wrote: > >>On July 21, 2004 at 12:21:38, James Swafford wrote: >> >>>First, my apologies for having the service down this long already. >>>I had hoped to have it back up by now. >>> >>>The search engine lives on a production web server that runs >>>a few websites and email for a small ISP that's jointly >>>owned by me and two others. One of our clients has been complaining >>>about getting 404 errors _immediately_ after requesting a page >>>from their site. At first I didn't think it was our problem, but >>>over time I became convinced it was. >>> >>>After scratching our heads for a while, we thought we'd stop the >>>search engine to see what, if any, effect that would have. The >>>problem struck infrequently enough that it wasn't immediately obvious; >>>we'd have to 'wait and see'. Two (three?) weeks later, there have >>>been no complaints. >>> >>>Andrei has done a great job with the search engine, but it does >>>have a (manageable) bug. Over time, it forks more and more client >>>processes and the load average goes through the roof. I think the >>>high load average was the cause of the 404 errors. >> >>While I hate to criticize Andrei's freely given work, infinite forking is _not_ >>a manageable bug. Tell him to fix it, and then (as he suggested) run it with >>nice 20 (making it the absolute lowest priority on the system). > >manageable != acceptable :) > > >I'm sure he'll fix it when he has time... he is very very busy these >days, and I can sympathize. Until then, I'm satisfied with the temporary >killall ccc_client, as long as it's on a non-critical box. > >-- >James > > > >> >>anthony
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.