dokucama

User Tools

Site Tools

Trace:
network_stuff:tcpdump

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revisionPrevious revision
Next revision		Previous revision
network_stuff:tcpdump [2021/07/27 15:48] – jotasandokunetwork_stuff:tcpdump [2023/11/02 14:38] (current) – external edit 127.0.0.1
Line 1: Line 1:
-**__TCPDUMP__** \\ +**__TCPDUMP NOTES__** \\ 
 +[[https://softbackbone.duckdns.org/doku.php?id=network_stuff:tcpnotes|tcp_notes]]\\
 To write the tcpdump output in pcap format. If we Add capital W, that means rotate: To write the tcpdump output in pcap format. If we Add capital W, that means rotate:
   tcpdump -ni eth0 -s0 -w /var/tmp/capture.pcap   tcpdump -ni eth0 -s0 -w /var/tmp/capture.pcap
   tcpdump -ni eth0 -w /var/tmp/trace -W 48 -G 1800 -C 100 -K    # 48 files, either every 1800 seconds (=30 minutes) or every 100 MB, -K don't verify checksum   tcpdump -ni eth0 -w /var/tmp/trace -W 48 -G 1800 -C 100 -K    # 48 files, either every 1800 seconds (=30 minutes) or every 100 MB, -K don't verify checksum
   tcpdump -nni bond1 -w /var/tmp/trace -W 1 -G 20 -C 100 -K   tcpdump -nni bond1 -w /var/tmp/trace -W 1 -G 20 -C 100 -K
 +  tcpdump -e -r sflow_2022_new.pcap   # to read an existing pcap (remove -e if not interested in ethernet headers)
  
 After capture, transfer example:  After capture, transfer example: 
  
-  gk# scp root@slingshot2.nj.grapeshot.co.uk:/var/tmp/capture.pcap /home/jaime_santos/Downloads/ +  gk# scp root@slingshot2.nj.mycompany1.co.uk:/var/tmp/capture.pcap /home/jaime_santos/Downloads/ 
-  hm# scp root@gatekeeper.grapeshot.co.uk:/home/jaime_santos/Downloads/capture.pcap capture-ams.pcap+  hm# scp root@gatekeeper.mycompany1.co.uk:/home/jaime_santos/Downloads/capture.pcap capture-ams.pcap
   remote# rm /var/tmp/capture.pcap   remote# rm /var/tmp/capture.pcap
   gk# rm /var/tmp/capture.pcap   gk# rm /var/tmp/capture.pcap
Line 24: Line 25:
   tcpdump -nni any portrange 60000-60100   tcpdump -nni any portrange 60000-60100
   tcpdump -vv -e -nni any ether proto 0x8809   # LACP   tcpdump -vv -e -nni any ether proto 0x8809   # LACP
 +  tcpdump -n -i eth0 -A -x dst port 443 and greater 100  # packet lenght
  
  
Line 93: Line 95:
  
 ---- ----
- +**LIBCAP RELATED UTILITIES**
-__**TCP NOTES**__ +
-\\ +
-This is announced during the tcp handshake:\\ +
- +
-  * MSS is announced (not really negotiated but just announced). +
-  * Window scaling is also announced. The default window size is 64kB which is far too small. That's way window scaling is ON 99% of the times. +
-  * SACK also announced in the 3 way handshake and also ON 99% of the times.  +
-    * SACK is like a lookeahead acknowledgement while we wait for slow Bytes to arrive. +
-    * Example: "If I received Bytes 1,2,3 5,6 but not 4, I acknowledge 3 and 'selectively acknowledge' 5 and 6." +
- +
-The ACK number for a packet is the packet's sequence number plus the data length. See this [[https://osqa-ask.wireshark.org/questions/43447/matching-tcp-packets-in-wireshark|Link]] +
- +
-\\ +
-  * The  congestion window is a **sender imposed** window that was implemented to avoid overrunning some routers in the middle of the network path. The **sender, with each segment sent, increases the congestion window slightly**, i.e. the sender will allow itself more outstanding sent data.  +
-    * You can't 'get' that value directly from the capture file, as **it is NOT ADVERTISED, it lives in the sender** +
-  * The **__receive window__** is the amount of data the **receiver can take at once without getting overwhelmed**. +
-    * Managed by the receiver, who sends out window sizes to the sender. The window sizes announce the number of bytes still free in the receiver buffer, i.e. the number of bytes the sender can still send without needing an acknowledgement from the receive +
-    * Is quite confusing because is the one that TRAVELS IN THE TCP HEADER. In books it appears as advertised window meaning that is **Advertised by the Receiver** +
-\\ +
- +
----- +
- +
-__TCP CONGESTION CONTROL ALGORITHMS__ (Reno, Cubic, Tahoe, more recently, BBR + Vegas, , Westwood) +
-[[https://medium.com/@atoonk/tcp-bbr-exploring-tcp-congestion-control-84c9c11dc3a9]] +
- +
-  * sysctl net.ipv4.tcp_congestion_control  # default is usually cubic or reno +
-  * sysctl net.ipv4.tcp_available_congestion_control  # list available suites +
-  * sysctl -w net.ipv4.tcp_congestion_control=bbr   # rebuilt suite so it takes latency as congestion, not packet losses (by google) +
- +
-To test performance: +
-  tc qdisc replace dev enp0s20f0 root netem loss 1.5% latency 70ms # introduces some latency and packet loss +
- +
- +
-WINDOWING:\\ +
-  * MSS and window scaling is negotiated at the beginning. Normally ~*128 +
- +
-When a port is not available and the connection is rejected, an ICMP unreachable message is sent and then a RST tcp packet +
- +
- +
----- +
-__TCP OPTIMIZATION__\\ +
-[[https://www.extrahop.com/company/blog/2016/tcp-nodelay-nagle-quickack-best-practices/]]+
 \\ \\
-  * **NAGLE**: Aim is to reduce the number of small packets sent over the network. You might want to fill up the truck instead of sending it just with one box, or not.. **Nagle's algorithm and delayed ACKs. Hence Nagle's algorithm is undesirable in highly interactive environments.**\\ +  tcprewrite -C --dlt=enet --enet-smac=3a:85:a1:b3:c3:84 --enet-dmac=c2:54:cb:f2:ee:c8  -i sflow_2022_new.pcap -o rwsflow_2022_new.pcap   # this is needed when the pcap we received is linux cooked (sometimes they miss the dst mac address so we add a dummy one 
-    * TCP_NODELAY socket option allows your network to bypass Nagle Delays by disabling Nagle's algorithm, and sending the data as soon as it's available +  tcprewrite -C --dstipmap=192.168.97.255/32:10.100.10.142/32 -i rwsflow_2022_new.pcap -o rwrwsflow_2022_new.pcap 
-  * **Delayed ACK**: is basically a bet taken by the destination betting 200 - 500 ms, that a new packet will arrive before the delayed ACK timer expires.  Nagle's algorithm effectively only allows one packet to be actively transporting on the network at any given time, this tends to hold back traffic due to the interactions between the Nagle's algorithm and delayed ACKs. +  tcpreplay -i eth0 --loop 0 xxx.pcap    
-    * To disable Delayed ACKs, use the TCP_QUICKACK socket option. +[[https://gist.github.com/niranjan-nagaraju/4532037]]
network_stuff/tcpdump.1627400922.txt.gz · Last modified: (external edit)

Page Tools

Except where otherwise noted, content on this wiki is licensed under the following license: CC Attribution-Share Alike 4.0 International
CC Attribution-Share Alike 4.0 International Donate Powered by PHP Valid HTML5 Valid CSS Driven by DokuWiki