
Mailing List Archive
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [tlug] Odd mmap() problem.
- Date: Wed, 20 Jul 2005 23:16:42 +1000 (EST)
- From: Jim Breen <Jim.Breen@example.com>
- Subject: Re: [tlug] Odd mmap() problem.
>> From: Botond Botyanszki <tlug@example.com>
>> On Mon, 18 Jul 2005 14:07:43 +1000 (EST)
>> Jim Breen <Jim.Breen@example.com> wrote:
>>
>> > My WWWJDIC server is barfing reporting "Out of space" (via strerror) as
>> > it fails an mmap() call. It occurs in the following situation:
>> >
>> > - just for one user (Michael) living in Nagoya;
>> I guess it means mmap returning with -1 and ENOMEM. I hardly see a
>> connection with the "symptoms". If this only happens for one user, it's
>> possible that you have a buffer overflow somewhere else in your C cgi
>> code. There are tools for detecting these, fedora core 4 comes with
>> builtin support in gcc+glibc for this.
The user forgot to tell me that his calls were coming from a
Javascript "button" with the biggest dictionary file selected. Once I
had exactly replicated his calls, it began failing for me too. Seems
the Solaris/Apache/CGI combo at Monash can't mmap more than about 48Mb
of files at a time. I have some workarounds, but I've switched my
attack^H^H^H^H^H^Hrequests to our sysadmins.
Nice that the problem doesn't happen at Linux sites.
Jim
--
Jim Breen http://www.csse.monash.edu.au/~jwb/
Clayton School of Information Technology, Tel: +61 3 9905 9554
Monash University, VIC 3800, Australia Fax: +61 3 9905 5146
(Monash Provider No. 00008C) ジム・ブリーン@モナシュ大学
Home |
Main Index |
Thread Index