On Tue, Mar 12, 2002 at 11:22:19AM -0500, Alan Eldridge wrote: >On Tue, Mar 12, 2002 at 09:59:42AM -0600, Claus wrote:
>>A small sample from the log files: >> >>2002-03-12 09:53:23 2 80.128.229.62 aldireview.niesens.com 80 "GET >>/logo.gif HTTP/1.0" 404 0 202 "http://aldireview.niesens.com/"; "Wget/1.7" >>2002-03-12 09:53:24 2 80.128.229.62 aldireview.niesens.com 80 "GET >>/index.en.html HTTP/1.0" 404 0 207 "http://aldireview.niesens.com/"; >>"Wget/1.7" 2002-03-12 09:53:24 2 80.128.229.62 aldireview.niesens.com 80 One other thing: the lack of a delay is the fault of the user invoking the program. Wget has plenty of options to handle delays between successful fetches and retries of failed fetches, including an incremental backoff feature, but if the user doesn't *use* them, then there's not much to do about it. If this is the bug I think it is, it won't go on for too long, depending on the user's memory size and swap space. Eventually, the machine will start swapping so much that the hits start coming farther apart due to disk thrashing, and either the wget will fail to get requested memory and die, the OS will kill the process for using too much memory (most Unix-like systems) or the OS will crash (that'd be my guess for Winbloze). That user's IP reverse-resolves to what looks like a dialup, so the hopes of reaching them to tell them to update their wget software are not very good, unfortunately. -- Alan Eldridge "Dave's not here, man."
