Wireshark mailing list archives

Re: Limiting amount of memory used to analyze TCP (HTTP) traffic. Questions about tcp dissector.


From: Jeff Morriss <jeff.morriss.ws () gmail com>
Date: Thu, 25 Feb 2016 17:58:32 -0500

On Thu, Feb 25, 2016 at 4:53 PM, Vitaly Repin <vitaly.repin () gmail com>
wrote:

Hello,

I am trying to understand how the Wireshark TCP dissector utilizes memory.


That's a good place to start but there's a lot of other stuff in Wireshark
that will use a lot of memory as time goes by.

On the off chance you haven't read it (at least a dozen times :-)) already:
Wireshark's generally not the best tool for doing long-term analysis for
exactly this reason (memory usage).

At one point I was quite interested in adding some kind of memory profiling
into Wireshark so we could see exactly where the memory was used at any
given point in time (e.g., frame_data's are using 20 Mbytes, TCP
reassembly's using 100 Mbytes, etc.) but I never came up with something
good.

Then I have taken a look into  tcpd->acked_table.  According to the
comment it "contains a tree containing all the various ta's keyed by
frame number".
I see that this list monothonically grows during the analysis. It is
expected behavior?


Assuming that 1) your TCP segments are being ACK'd and 2) you have TCP
sequence analysis enabled (it is by default) then yes, it is expected.


Any ideas how I can decrease memory consumption (even for the price of
not being able to analyze the whole TCP session if it contains huge
amount of data)?


Have you gone through the suggestions in the Wiki's OutOfMemory article?
It has some pointers to things that will limit the memory usage.

https://wiki.wireshark.org/KnownBugs/OutOfMemory
___________________________________________________________________________
Sent via:    Wireshark-dev mailing list <wireshark-dev () wireshark org>
Archives:    https://www.wireshark.org/lists/wireshark-dev
Unsubscribe: https://wireshark.org/mailman/options/wireshark-dev
             mailto:wireshark-dev-request () wireshark org?subject=unsubscribe

Current thread: