I used the GET method, and I now know the search term used. The user searched for the word "recall" . Does anyone know why this would cause a problem, or how to avoid this happening again. I have not tried reproducing the error yet, but may risk trying later on. > I am running htdig release 3.2.0b3 on linux kernel 2.4.2-2smp > on a production web server. > > Htsearch normally works fine, but on a couple of occasions it has > generated over 2 million lines of the following error in the error logs: > > WordKey::Compare: key length for a or b < info.num_length > > Unfortunately I do not know what the user entered as search terms when > the errors occured, but I do know that the time of these 2 million + > entries in the error log coincides with an htsearch request in the > access log. > > If anyone knows what causes this or how to rectify the problem then > I would really appreciate a response. Gilles Detillieux wrote: > According to James Dove: > > I am running htdig release 3.2.0b3 on linux kernel 2.4.2-2smp > > on a production web server. > > > > Htsearch normally works fine, but on a couple of occasions it has > > generated over 2 million lines of the following error in the error logs: > > > > WordKey::Compare: key length for a or b < info.num_length > > > > Unfortunately I do not know what the user entered as search terms when > > the errors occured, but I do know that the time of these 2 million + > > entries in the error log coincides with an htsearch request in the > > access log. > > > > If anyone knows what causes this or how to rectify the problem then > > I would really appreciate a response. > > If you use the GET method in your search form, rather than POST, you'll > be able to see in your access log what search terms were used, so you > can reproduce the problem. > > However, you may want to try out the htdig-3.2.0b4-090301.tar.gz snapshot > from ftp://ftp.htdig.org/pub/htdig/snapshots/, to see if it helps with > this problem. I don't know if any changes addressed this problem directly > (it's been reported before, but never consistently enough for us to pin > it down), but it has a large number of other bug fixes, some of which > could possibly indirectly prevent this. Let us know if the problem > persists with the latest snapshot, and see if you can come up with a > reproduceable test case. > > -- > Gilles R. Detillieux E-mail: <[EMAIL PROTECTED]> > Spinal Cord Research Centre WWW: http://www.scrc.umanitoba.ca/~grdetil > Dept. Physiology, U. of Manitoba Phone: (204)789-3766 > Winnipeg, MB R3E 3J7 (Canada) Fax: (204)789-3930 _______________________________________________ htdig-general mailing list <[EMAIL PROTECTED]> To unsubscribe, send a message to <[EMAIL PROTECTED]> with a subject of unsubscribe FAQ: http://htdig.sourceforge.net/FAQ.html
