[Snort-devel] SNORT has memory leak on Linux Red hat 9
kumarm at ...2330...
Wed Feb 25 15:19:07 EST 2004
Thanks for your reply
Yes,exactly this is what very odd is that even I kill the SNORT process,memory are not getting released. This is really strange. But,this is what exactly happening.
I have 4 GB of physical memory. I stopped all the process and waited for 3-4 hours.It remains at 345MB. As soon as I start the snort process,it starts climbing and other thing is that it climbes very fast. Within 2-3 hours, it eats up 95% of the memory. It goes upto 3.5 GB. When I kill it,I thought it will release the memory,but it's not.
Again,this is happening on Red hat 9 and also on application service 2.4.9 Linux.
From: Ian Macdonald [mailto:ism at ...2371...]
Sent: Wednesday, February 25, 2004 6:06 PM
To: Kumar, Manoj
Cc: Jeremy Hewlett; snort-users at lists.sourceforge.net;
snort-devel at lists.sourceforge.net; snort-announce at lists.sourceforge.net
Subject: Re: [Snort-devel] SNORT has memory leak on Linux Red hat 9
The first thing that seems odd is that the memory doesn't free up after
killing the process. Normally all memory would be released on application
termination. When you say memory doesn't free what item are you looking
One thing you might want to do is try killing other applications to see if
they are the ones that are stealing the memory. You may even want to
remove the loaded modules one by one incase the memory leak is in the
network driver module.
The only other thing of the top of my head is that machine is swapping so
much that it takes time for the OS to swap out the memory from disk to
allow it to be released?
Have you tried it on a different OS or Kernel?
> Hello everybody,
> I am running SNORT ver 2.1.0 to capture data from my giga bit network on
> RedHat Linux 9 where SNORT is capturing 100MB of data per minute (Lots of
> data). Problem is that memory usage keeps going as long as SNORT is
> WORST thing is that even if you kill the SNORT process, it doesn't release
> the memory. Memory usage remains as it is.
> Would you guys please help me out? Why SNORT is behaving like this and
> anybody has noticed this problem?
> -----Original Message-----
> From: Jeremy Hewlett [ mailto:jh at ...402...]
> Sent: Wednesday, February 25, 2004 4:41 PM
> To: snort-users at lists.sourceforge.net
> Cc: snort-devel at lists.sourceforge.net;
> snort-announce at lists.sourceforge.net
> Subject: [Snort-devel] Snort 2.1.1 final is available!
> Snort 2.1.1 is now available - Thanks everyone who installed RC1 and
> tried it out! The differences between RC1 and final are minor, and
> * Documentation updates and fixes by JP Vossen, Felipe Franciosi, and
> Drew Smith
> * Compiles on Tru64 now - thanks Hari Gopal and Darryl Cook.
> * libintsnort.a is no longer included in compile routine (this is the
> Solaris "ar" problem some people have had)
> * Snort templates have been updated
> * Fixed issue with CSV not displaying its output correctly - thanks
> Bill Guyton and Alan Milligan for your fixes.
> * Fixed Flow-Portscan alert-mode bug where only one alert would get
> generated. Thanks Kevin Amorin for pointing out the problem and
> testing the fix.
> * Minor Makefile fix for "unexpected end of line" at the verstuff.pl
> line when not using GNU "make" on Solaris - Thanks for the report,
> Chad Kreimendahl.
> * Removed escaping of '%' and '_' characters in MySQL (thanks
> Kristofer Karas).
> For further info on changes, please review the ChangeLog and
> RELEASE.NOTES, which can be found in the parent directory of the snort
> Happy Snorting,
> The Snort Team
> SF.Net is sponsored by: Speed Start Your Linux Apps Now.
> Build and deploy apps & Web services for Linux with
> a free DVD software kit from IBM. Click Now!
> Snort-devel mailing list
> Snort-devel at lists.sourceforge.net
More information about the Snort-devel