Archive for the ‘Networking’ Category

Why and How to Increase and decrease txqueuelen (Transmit Queue Lenght) in GNU / Linux

Friday, September 22nd, 2017

linuxnetworkingchangetxquelenincreasedecreasefornetworklatencyimprove

In GNU / Linux network routers and sometimes even home PCs it is mostly helpful to play with a TCP / IP stack Network Interface Card (NIC) value called (txqueuelen) in order to make the Network interface (Ethernet device) to work better with the type of network that is connected to it. On a slower LAN or Internet connections txqueuelen is better to be decreased whether on a high speed connections raising it will increase the performance of network traffic to perform large homogenous network transfers.

Therefore on a small private networks in many occasions (though it depends on the served services and exact type of network) it is more useful to decrease the txqueuelen Linux value whether on a High latency networks increasing it will reap a great benefits for your ISP or Hosting routers.

So what is txqueuelen value instructs the kernel on the largeness of transmit queue of the network interface device.
 

1. Decrease the number of txqueuelen in GNU / Linux on a slower networks (ADSL routers), Mobile networks etc to improve network latency


For a slower router devices be it WI-FI routers such as D-Link or any cheap chineese Wifi or LAN or Telephone or ISP compnay networks that provide internet via ADSL routers that run Linux or has access to Linux console with available access to ifconfig command or have an iproute2 Linux package installed (most of them have) it is very helpful to reduce its size for a smaller values in order to guarantee a network high latency.

Therefore reducing the txquelen value to a number like 200 for ADSL provided internet can benefit you.
Assuming that you know the NIC name in Linux the first one is usually eth0 in order to reduce this value to 200 issue:

 

ifconfig eth0 txqueuelen length
ifconfig eth0 txqueuelen 200

 

To do the same with iproute2 (ip) command in case if the router is missing ifconfig or you just prefer to use iproute2 (advanced and newer bunch of commands that improved the Linux networking functionalities thanks to Russian Alexei Kuznecov 15 years ago or so) run:

 

ip link set eth0 txqueuelen length
ip link set eth0 txqueuelen 200

 

To make it permanent you can either create a brand new rc. script lets say /etc/init.d/rc.queuelenght and add the commands or use the good old /etc/rc.local to make the commands get loaded on every GNU / linux or router boot time.

Above reducement will make your Network TCP stack more responsive.

2. Increase the number of txquelen in GNU / Linux to improve the network latency on a company NAT or other ISP routers

Depending on the scale of Internet bandwidth that is coming inside your router you should decide how much you would like to increase the txqueuelen value.
Generally speaking txqueuelen value could be played with on a Large IN / OUT Company Routers with values between 1000 and 20000

Most GNU / Linux distributions comes preconfigured as a default a value of 1000, so lets say you would like to raise the value to 5000 to improve network transfer latency, here is how to do it via ifconfig:

 

ifconfig eth0 txqueuelen length
ifconfig eth0 txqueuelen 5000

 

With IPRoute2 analogously run:

 

ip link set eth0 txqueuelen length
ip link set eth0 txqueuelen 5000

 

Of course the best way to figure out what will be the best value for your case is to experiment a little bit with the value and use some kind of speedtest (test bandwitdh service – this one is mine) from the many available online.
You have to consider the type of hardware of Server and most importantly the Network Card hardware (which vendor and what are the recommended ones from the vendor).

If the router is some Linux distribution lets say a Debian / Ubuntu Linux etc other deb based one, it is better to make the necessery permanent changes to make them be loaded on system boot not via /etc/rc.local but by adding to file /etc/network/interfaces:

 

vim /etc/network/interfaces

 

up ifconfig $IFACE txqueuelen 8000

 

Note that as network traffic processing depends on CPU power, decreasing the txqueuelen will add a little bit of extra load to the Central Processing Unit.


In RHEL / CentOS Linux and SuSE server a good way to increase txqueuelen network latency is by running:

 

# cat <> /etc/rules.d/71-net-txqueuelen.rules
SUBSYSTEM==”net”, ACTION==”add”, KERNEL==”eth*”, ATTR{tx_queue_len}=”10000″
EOF

 

 

udevadm trigger

 

I've red on the internet that a lot of people are reporting a very good results with a txqueuelen setting of 2000 as this value is said to give good results and not hammering the CPU too hardly, so it might be a good idea if you're experimenting with values to find out which one suits you better is to start with this one, e.g. run:
 

ifconfig eth0 txqueuelen 2000
echo 2000 > /proc/sys/net/core/netdev_max_backlog


Well of course keep in mind that sometimes increasing the txqueuelen to too large values will make you an easier target for Denial of Service Attack attempts, though most modern computer routers should be able to behave good even with the highest value like 10000.
 

 

 

Share this on

Mount remote Linux SSHFS Filesystem harddisk on Windows Explorer SWISH SSHFS file mounter and a short evaluation on what is available to copy files to SSHFS from Windows PC

Monday, February 22nd, 2016

swish-mount-and-copy-files-from-windows-to-linux-via-sshfs-mount

I'm forced to use Windows on my workbook and I found it really irritating, that I can't easily share files in a DropBox, Google Drive, MS OneDrive, Amazon Storage or other cloud-storage free remote service. etc.
I don't want to use DropBox like non self-hosted Data storage because I want to keep my data private and therefore the only and best option for me was to make possible share my Linux harddisk storage
dir remotely to the Windows notebook.

I didn't wanted to setup some complex environment such as Samba Share Server (which used to be often a common option to share file from Linux server to Windows), neither wanted to bother with  installing FTP service and bother with FTP clients, or configuring some other complex stuff such as WebDav – which BTW is an accepted and heavily used solution across corporate clients to access read / write files on a remote Linux servers.
Hence, I made a quick research what else besides could be used to easily share files and data from Windows PC / notebook to a home brew or professional hosting Linux server.

It turned out, there are few of softwares that gives a similar possibility for a home lan small network Linux / Windows hybrid network users such, here is few of the many:

  • SyncThingSyncthing is an open-source file synchronization client/server application, written in Go, implementing its own, equally free Block Exchange Protocol. The source code's content-management repository is hosted on GitHub

     

     

     

     

     

    syncthing-logo

  • OwnCloud – ownCloud provides universal access to your files via the web, your computer or your mobile devices

     

     

     

     

     

    owncloud-logo

  • Seafile – Seafile is a file hosting software system. Files are stored on a central server and can be synchronized with personal computers and mobile devices via the Seafile client. Files can also be accessed via the server's web interface


seafile-client-in-browser

I've checked all of them and give a quick try of Syncthing which is really easy to start, just download the binary launch it and configure it under https://Localhost:8385 URL from a browser on the Linux server.
Syncthing seemed to be nice and easy to configure solution to be able to Sync files between Server A (Windows) and Server B(Linux) and guess many would enjoy it, if you want to give it a try you can follow this short install syncthing article.
However what I didsliked in both SyncThing and OwnCloud and Seafile and all of the other Sync file solutions was, they only supported synchronization via web and didn't seemed to have a Windows Explorer integration and did required
the server to run more services, posing another security hole in the system as such third party softwares are not easily to update and maintain.

Because of that finally after rethinking about some other ways to copy files to a locally mounted Sync directory from the Linux server, I've decided to give SSHFS a try. Mounting SSHFS between two Linux / UNIX hosts is
quite easy task with SSHFS tool

In Windows however the only way I know to transfer files to Linux via SSHFS was with WinSCP client and other SCP clients as well as the experimental:

As well as few others such as ExpandDrive, Netdrive, Dokan SSHFS (mirrored for download here)
I should say that I first decided to try copying few dozen of Gigabyte movies, text, books etc. using WinSCP direct connection, but after getting a couple of timeouts I was tired of WinSCP and decided to look for better way to copy to remote Linux SSHFS.
However the best solution I found after a bit of extensive turned to be:

SWISH – Easy SFTP for Windows

Swish is very straight forward to configure compared to all of them you download the .exe which as of time of writting is at version 0.8.0 install on the PC and right in My Computer you will get a New Device called Swish next to your local and remote drives C:/ D:/ , USBs etc.

swish-new-device-to-appear-in-my-computer-to-mount-sshfs

As you see in below screenshot two new non-standard buttons will Appear in Windows Explorer that lets you configure SWISH

windows-mount-sshfs-swish-add-sftp-connection-button-screenshot

Next and final step before you have the SSHFS remote Linux filesystem visible on Windows Xp / 7 / 8 / 10 is to fill in remote Linux hostname address (or even better fill in IP to get rid of possible DNS issues), UserName (UserID) and Direcory to mount.

swish-new-fill-in-dialog-to-make-new-linux-sshfs-mount-directory-possible-on-windows

Then you will see the SSHFS moutned:

swish-sshfs-mounted-on-windows

You will be asked to accept the SSH host-key as it used to be unknown so far

swish-mount-sshfs-partition-on-windows-from-remote-linux-accept-key

That's it now you will see straight into Windows Explorer the remote Linux SSHFS mounted:

remote-sshfs-linux-filesystem-mounted-in-windows-explorer-with-swish

Once setupped a Swish connection to copy files directly to it you can use the Send to Embedded Windows dialog, as in below screenshot

swish-send-to-files-windows-screenshot

The only 3 problem with SWISH are:

1. It doesn't support Save password, so on every Windows PC reboot when you want to connect to remote Linux SSHFS, you will have to retype remote login user pass.
Fron security stand point this is not such a bad thing, but it is a bit irritating to everytime type the password without an option to save permanently.
The good thing here is you can use Launch Key Agent
as visible in above screenshot and set in Putty Key Agent your remote host SSH key so the passwordless login will work without any authentication at
all however, this might open a security hole if your Win PC gets infected by virus, which might delete something on remote mounted SSHFS filesystem so I personally prefer to retype password on every boot.

2. it is a bit slow so if you're planning to Transfer large amounts of Data as hundreds of megabytes, expect a very slow transfer rate, even in a Local  10Mbit Network to transfer 20 – 30 GB of data, it took me about 2-3 hours or so.
SWISH is not actively supported and it doesn't have new release since 20th of June 2013, but for the general work I need it is just perfect, as I don't tent to be Synchronizing Dozens of Gigabytes all the time between my notebook PC and the Linux server.

3. If you don't use the established mounted connection for a while or your computer goes to sleep mode after recovering your connection to remote Linux HDD if opened in Windows File Explorer will probably be dead and you will have to re-enable it.

For Mac OS X users who want to mount / attach remote directory from a Linux partitions should look in fuguA Mac OS X SFTP, SCP and SSH Frontend

I'll be glad to hear from people on other good ways to achieve same results as with SWISH but have a better copy speed while using SSHFS.

Share this on

Apache Webserver disable hostnamelookups “HostnameLookups off” for minor performance increase

Friday, February 12th, 2016

apache-disable-dns-lookups-for-speed-hostnamelookups-off-directive-building-scalable-php-applications

If you don't much care about logging in logs from which domain / hostnames requests to webserver originate and you want to boost up the Apache Webserver performance a bit especially on a heavy loaded Websites, where no need for stuff like Webalizer, Awstats etc. , e.g. you're using GoogleAnalytics to already track requests (beware as sometimes GoogleAnalytics could be missing requests to your webserver, so having some kind of LogAnalyzer software on server is always a plus). But anyways accepting that many of us already trust GoogleAnalytitcs.


Then a great tuning option to use in default domain configuration or in multiple VirtualHosts config is:

HostnameLookups off

If you want to make the HostnameLookups off as a default behaviour to all your virtualhosts on  Debian / Ubuntu / CentOS / SuSE / RHEL distro virtualhosts add either to default config /etc/apache2/sites-enabled/000-default (on Deb based Linuxes) or (on RPM based ones), add directive to /etc/httpd/httpd.conf

For self-hosted websites (if run your own small hosting) or for a home situated webservers with up to 20-50 websites it is also a useful optimization tip to include in /etc/hosts file all the IPs of sites with respective domain names following the normal syntax of /etc/hosts, e.g. in my own /etc/hosts, I have stuff like:
 

pcfreak:~$ cat /etc/hosts
127.0.0.1 localhost localhost.localdomain
127.0.1.1 pcfreak.pc-freak.net pcfreak mail.pc-freak.net
192.168.0.14 new-pcfreak
219.22.88.70 fw
212.36.0.70 ftp.bg.debian.org
212.211.132.32 security.debian.org
83.228.93.76 pcfreak.biz pc-freak.net www.pc-freak.net
# for wordpress plugins
216.58.209.3 gstatic.com
91.225.248.129 www.linkedin.com
74.50.119.198 www.blogtopsites.com
94.31.29.40 static.addtoany.com
216.58.209.202 fonts.googleapis.com
216.58.209.14 www.google-analytics.com
216.58.209.14 feeds.feedburner.com
93.184.220.241 wprp.zemanta.com
199.30.80.32 stumbleupon.com
156.154.168.17 stumbleupon.com
2.18.89.251 platform.linkedin.com
# The following lines are desirable for IPv6 capable hosts

# … etc. put IPs and hostnames following above syntax


As you see from above commented section for wordpress plugins, I've included some common websites used by WordPress enabled plugins to prevent my own hosting server to query DNS server every time. The normal way the Linux / Unix works is it first checks in /etc/hosts and only if the hostname is not defined there then it queries the DNS caching server in my case this is a local DJBDNS cache server, however defining the hosts in /etc/hosts saves a lot of milisecons on every request and often if multiple hosts are defined could save (decrease site opening for end users) with seconds.


Well now use some website speed testing plugin like Yslow, Firebug Fiddler or HTTPWatch

 

Share this on

How to mount NFS network filesystem to remote server via /etc/fstab on Linux

Friday, January 29th, 2016

mount-nfs-in-linux-via--etc-fstab-howto-mount-remote-partitions-from-application-server-to-storage-server
If you have a server topology part of a project where 3 (A, B, C) servers need to be used to deliver a service (one with application server such as Jboss / Tomcat / Apache, second just as a Storage Server holding a dozens of LVM-ed SSD hard drives and an Oracle database backend to provide data about the project) and you need to access server A (application server) to server B (the Storage "monster") one common solution is to use NFS (Network FileSystem) Mount. 
NFS mount is considered already a bit of obsoleted technology as it is generally considered unsecre, however if SSHFS mount is not required due to initial design decision or because both servers A and B are staying in a serious firewalled (DMZ) dedicated networ then NTS should be a good choice.
Of course to use NFS mount should always be a carefully selected Environment Architect decision so remote NFS mount, imply  that both servers are connected via a high-speed gigabyte network, e.g. network performance is calculated to be enough for application A <-> to network storage B two sides communication not to cause delays for systems end Users.

To test whether the NFS server B mount is possible on the application server A, type something like:

 

mount -t nfs -o soft,timeo=900,retrans=3,vers=3, proto=tcp remotenfsserver-host:/home/nfs-mount-data /mnt/nfs-mount-point


If the mount is fine to make the mount permanent on application server host A (in case of server reboot), add to /etc/fstab end of file, following:

1.2.3.4:/application/local-application-dir-to-mount /application/remote-application-dir-to-mount nfs   rw,bg,nolock,vers=3,tcp,timeo=600,rsize=32768,wsize=32768,hard,intr 1 2


If the NTFS server has a hostname you can also type hostname instead of above example sample IP 1.2.3.4, this is however not recommended as this might cause in case of DNS or Domain problems.
If you want to mount with hostname (in case if storage server IP is being commonly changed due to auto-selection from a DHCP server):

server-hostA:/application/local-application-dir-to-mount /application/remote-application-dir-to-mount nfs   rw,bg,nolock,vers=3,tcp,timeo=600,rsize=32768,wsize=32768,hard,intr 1 2

In above example you need to have the /application/local-application-dir-to-mount (dir where remote NFS folder will be mounted on server A) as well as the /application/remote-application-dir-to-mount
Also on server Storage B server, you have to have running NFS server with firewall accessibility from server A working.

The timeou=600 (is defined in) order to make the timeout for remote NFS accessibility 1 hour in order to escape mount failures if there is some minutes network failure between server A and server B, the rsize and wsize
should be fine tuned according to the files that are being red from remote NFS server and the network speed between the two in the example are due to environment architecture (e.g. to reflect the type of files that are being transferred by the 2)
and the remote NFS server running version and the Linux kernel versions, these settings are for Linux kernel branch 2.6.18.x which as of time of writting this article is obsolete, so if you want to use the settings check for your kernel version and
NTFS and google and experiment.

Anyways, if you're not sure about wsize and and rise, its perfectly safe to omit these 2 values if you're not familiar to it.

To finally check the NFS mount is fine,  grep it:

 

# mount|grep -i nfs
sunrpc on /var/lib/nfs/rpc_pipefs type rpc_pipefs (rw)
server-hostA:/application/remote-application-dir-to-mount on /application/remote-application-dir-to-mount type nfs (rw,bg,nolock,nfsvers=3,tcp,timeo=600,rsize=32768,wsize=32768,hard,intr,addr=1.2.3.4)


That's all enjoy 🙂

 

 

Share this on

Secure your work PC internet traffic using SSH Dynamic Tunnel as Proxy to get around Corporate Spy Proxy and Site Filtering

Friday, March 20th, 2015

use-ssh-dynamic-tunnel-as-socks5-proxy-to-get-around-corporate-website-filtering-restrictions

If you work for some huge corporations such as IBM / Sony / Toshiba / Concentrix / HP etc. and you're using a Windows Work Computer (notebook), pre-installed with a custom Company software which is by default configured to use a Proxy Server for all your Browsing activities and at a certain point you start being filtered some of the websites you love to visit so much because of some Corporate policies (limitations) at some filtered sites you will start getting empty pages or some   nasty filtering messages.

Even if you don't get a filtering message but you know all your Company Internal Network traffic is proxified for the sake of keeping your personal (privacy) high stop browsing using company's default proxy, because all your access requests (passwords) and queries to the internet are probably logged for later (review) in case if you enter the company's paragraph of "non-compliant employee".
If you fail on time to get around the default set "Corporate Proxy", sooner or later you will start getting filtering messages to some of the regular websites you use daily, as I did today while trying to open my personal blog (to check if there are new user comments):

Your request was denied because of its content categorization: "Hacking;Malicious Sources/Malnets;Religion"
For assistance, contact your network support team.

Screenshot of above message from today here

You see this guys or automated Proxy filter became so prudent that my site was filtered because it contains some Proof of Concept (PoC) security tools and content related to Christian (Faith) Religion. I guess its the time to think seriously is there a censorship in large corporations and how far could censorship go and if such censorship so easily adopted in large companies wouldn't same happen also on a backbone ISP level in short future??
If today my site is being filtered out to be unable to open from a corporation network because it contains "Religious" contain I would not be surprised if tomorrow, I've been prohibited to confess publicly my faith in salvation power of the Cross of our Lord Jesus Christ or even already in a blacklist because I'm trying to be a dedicated Orthodox Christian …
The fact that Religion is already perceived in same light as Hacking and Malicious Source or Malnet bots is also very eloquent and shows how very big part of people nowdays (including the person that added my site to this proxy filtering rules) think of religion and in what bad state our society and understanding of freedom and respect for others went.

Obviously it is time to react to this censorship and stop the evil corporation from spying on your traffic and logging all that matches there "kilometer long" prohibited sites filter lists. There are few ways to do that and the most straight forward is to set-up and use a Own Proxy server such as Privoxy / Polipo or Squid Proxy, however the proxy method requires that your company local network doesn't have too strick (restrictive) firewall rules (e.g. you need some port opened to the Internet such as 8080, 3128, 8118, 1080 standard port for (socks) etc.

As many companies are too restrictive in their outbound firewall rules and you might be in situation like with me where Browsers such as Internet Explorer / Opera / Firefox and Chrome are configured to use by default company proxy host (autocache.proxy-ur-company.hp.com:80) (with a custom Proxy PAC file filtering out a whole ranges of useful domains and IPs) and only allowed firewall access outside of local corporate network in on port 22 (for outside ssh session purposes) only.

Then your best way to get across such restrictive network configuration is to run your own home Linux / BSD / Windows server with opensshd installed and use OpenSSH protocol Dynamic Tunneling (Proxy socks5 like) capabilities to tunnel all your favourite Web Browser Traffic (lets say Firefox's) through your remote-home-host.com:22.

 


In short once you have installed plink.exe on your PC run manually from command line (cmd.exe)

 

plink.exe -ssh UserName@remote-home-host.com -P 22 -pw Secret_Password -D 127.0.0.1:8080 -N


For people who use MobaXTerm it is even easier as there is an integrated SSH tunneling input interface which can be used to create the SSH tunnel.

To have a quick way to Enable SSH Dynamic Tunnel button on your Desktop make a SymLink to Plink with Target below command line:

web-tunnel-maker-with-plink-win-ssh-connection-tool-screenshot-on-ms-windows-7

  • If from Linux / *BSD / Mac OS host to create Dynamic SSH Tunnel to your remote home SSH server host run in a Terminal
     

ssh -D 8080 Username@remote-home-host.com


To start tunneling all your Web traffic via just created Dynamic SSH Tunnel to host remote-home-host.com, just set in browser's proxy options to use as proxy socks5 – localhost:8080

Secure-your-work-PC-notebook-internet-traffic-using-SSH-Dynamic-Tunnel-as-Proxy

To test whether your traffic is going to the Internet from remote-home-host.com open in just set proxy browser www.myip.ru .
You should see your home SSH server IP as IP which made the request to www.myip.ru.

Share this on

Linux: How to see / change supported network bandwidth of NIC interface and get various eth network statistics with ethtool

Monday, January 19th, 2015

linux-how-to-see-change-supported-network-bandwidth-of-NIC-interface-and-view-network-statistics
If you're a novice Linux sysadmin and inherited some dedicated servers without any documentation and hence on of the first things you have to do to start a new server documentation is to check the supported TCP/IP network speed of servers Network (ethernet) Interfaces. On Linux this is very easy task to verify the speed of LAN card supported Local / Internet traffic install ethtool (if not already preseont on the servers) – assuming you're dealing with Debian / Ubuntu Linux servers.

1. Install ethtool on Deb and RPM based distros

dedi-server1:~# apt-cache show ethtool|grep -i desc -A 3
Description: display or change Ethernet device settings
 ethtool can be used to query and change settings such as speed, auto-
 negotiation and checksum offload on many network devices, especially
 Ethernet devices.

dedi-server1:~# apt-get install –yes ethtool
..

ethtool should be installed by default on CentOS / Fedora / RHEL and  syntax is same like on Debs. If you happen to miss ethtool on any (SuSE) / RedHat / RPM based distro install it with yum

[root@centos:~] # yum -y install ethtool


2. Get ethernet configurations

To check the current eth0 / eth1 / ethX network (Speed / Duplex) and other network related configuration configuration:
 

dedi-server5:~# ethtool eth0

Settings for eth0:
        Supported ports: [ TP ]
        Supported link modes:   10baseT/Half 10baseT/Full
                                100baseT/Half 100baseT/Full
                                1000baseT/Full
        Supports auto-negotiation: Yes
        Advertised link modes:  10baseT/Half 10baseT/Full
                                100baseT/Half 100baseT/Full
                                1000baseT/Full

        Advertised pause frame use: No
        Advertised auto-negotiation: Yes
        Speed: 1000Mb/s
        Duplex: Full
        Port: Twisted Pair
        PHYAD: 1
        Transceiver: internal
        Auto-negotiation: on
        MDI-X: off
        Supports Wake-on: pumbag
        Wake-on: g
        Current message level: 0x00000001 (1)
        Link detected: yes

Having a NIC configured to act as Duplex is very important as Duplex communication enables LAN card to communicate both sides (Sent / Receive) packets simultaneously.

full-duplex-half-duplex-explained-picture

Probably most interesting parameters for most admins are the ones that are telling whether the NIC UpLink is 10megabyte / 100 megabyte or 1Gigabyte as well as supported Receive / Send ( Transfer ) speeds of LAN, a common useful ethtool admin use to just show current LAN ethernet interface speed:

server-admin1:~# ethtool eth0 |grep -i speed
        Speed: 1000Mb/s

 

To get info about NIC (kernel module / driver) used with ethtool:

dedi-server3:~# ethtool -i eth0 driver: e1000e
version: 1.2.20-k2
firmware-version: 1.8-0
bus-info: 0000:06:00.0

3. Make LAN Card blink to recognize eth is mapped to which Physical LAN

Besides that ethtool has many other useful use cases, for example if you have a server with 5 lan or more LAN cards, but you're not sure to which of all different EthX interfaces correspond, a very useful thing is to make eth0, eth1, eth2, eth3, etc. blink for 5 seconds in order to identify which static IP is binded physically to which NIC , here is how:

ethtool -p eth0 5


Then you can follow the procedure for any interface on the server and map them with a sticker 🙂

Ethtool is also useful for getting "deep" (thorough) statistics on Server LAN cards, this could be useful to identify sometimes hard to determine broadcast flood attacks:
 

4. Get network statistics with ethtool for interfaces
 

dedi-server5:~# ethtool -S eth0|grep -vw 0
NIC statistics:
     rx_packets: 6196644448
     tx_packets: 7197385158
     rx_bytes: 2038559235701
     tx_bytes: 8281206569250
     rx_broadcast: 357508947
     tx_broadcast: 172
     rx_multicast: 34731963
     tx_multicast: 20
     rx_errors: 115
     multicast: 34731963
     rx_length_errors: 115
     rx_no_buffer_count: 26391
     rx_missed_errors: 10059
     tx_timeout_count: 3
     tx_restart_queue: 2590
     rx_short_length_errors: 115
     tx_tcp_seg_good: 964136993
     rx_long_byte_count: 2038559235701
     rx_csum_offload_good: 5824813965
     rx_csum_offload_errors: 42186
     rx_smbus: 383640020

5. Turn on Auto Negotiation and change NIC set speed to 10 / 100 / 1000 Mb/s

Auto-negotiation is important as an ethernet procedure by which two communication devices (2 network cards) choose common transmission parameters such as speed, duplex mode, and flow control in order to achieve maximum transmission speed over the network. On 1000BASE-T basednetworks the standard is a mandatory. There is also backward compatability for older 10BASE-T Networks.

a) To raise up NIC to use 1000 Mb/s in case if the bandwidth was raised to 1Gb/s but NIC settings were not changed:

dedi-server1:~# ethtool -s eth0 speed 1000 duplex half autoneg off


b) In case if LAN speed has to be reduced for some weird reason to 10 / 100Mb/s

 

dedi-server1:~# ethtool -s eth0 speed 10 duplex half autoneg off

dedi-server1:~# ethtool -s eth0 speed 100 duplex half autoneg off

c) To enable disable NIC Autonegotiation:

dedi-server1:~# ethtool -s eth0 autoneg on


6. Change Speed / Duplex settings to load on boot

a) Set Network to Duplex on Fedora / CentOS etc.

Quickest way to do it is of course to use /etc/rc.local. If you want to do it following distribution logic on CentOS / RHEL Linux:

Add to /etc/sysconfig/network-scripts/ifcfg-eth0

vim /etc/sysconfig/network/-scripts/ifcfg-eth0

ETHTOOL_OPTS="speed 1000 duplex full autoneg off"

To load the new settings restart networking (be careful to have physical access to server if something goes wrong 🙂 )

service network restart

b) Change network speed / duplex setting on Debian / Ubuntu Linux

Add at the end of /etc/network/interfaces

vim /etc/network/interfaces

post-up ethtool -s eth0 speed 100 duplex full autoneg off

7. Tune NIC ring buffers

dedi-server1:~# ethtool -g eth0

Ring parameters for eth0:
Pre-set maximums:
RX:             4096
RX Mini:        0
RX Jumbo:       0
TX:             4096
Current hardware settings:
RX:             256
RX Mini:        0
RX Jumbo:       0
TX:             256

As you can see the default setting of RX (receive) buffer size is low 256 and on busy servers with high traffic loads, depending on the hardware NIC vendor this RX buffer size varies.
Through increasing the Rx/Tx ring buffer size , you can decrease the probability of discarding packets in the NIC during a scheduling delay.
A change in rx buffer ring requires NIC restart so (be careful not to loose connection to remote server), be sure to have iLO access to it.

Here is how to raise Rx ring buffer size 4 times from default value:

ethtool -G eth0 rx 4096 tx 4069

Share this on

Creating Dynamic SSH Tunnel on Windows with Plink – Scriptable SSH Tunnels on Windows

Friday, July 4th, 2014

creating-ssh-tunnel-on-windows-with-plink-ssh-tunnel-diagram-tunnel-email-traffic

In my earlier articles I've explained about Creating SSH Tunnels on Linux and BSD and how to create SSH Tunnels with Putty Terminal client on Windows.

Creating SSH Tunnels is a precious knowledge every advanced computer user, a system administrator or security expert should be well aware of.

Probably still there are plenty of people for which the SSH Tunnelling is something never heard of or even if heard it is a mythical term covered by ancient shadows 🙂

So What is an SSH Tunnel? – SSH Tunnels shortly explained.

A SSH tunnel consists of an encrypted tunnel created through a SSH protocol
connection
.

An SSH tunnel can be used to transfer unencrypted traffic over a
network through an encrypted channel
between two hosts (host A and Host B) usually using remote and local port.

Requirements for SSH tunnel

Where Host B has to be running a version of OpenSSH server on some reachable port from Host B, (lets say 22).
Host B OpenSSH server has to be with (enabled X11Forwarding X11Forwarding yes in /etc/ssh/sshd_config) and Host A needs to have some SSH client supporting port forwarding (ssh command on Linux – part of (openssh-client package) and on Windows – Putty / Plink or any other of the many available ssh tunneling clients).

Probably most common SSH Tunnel use is to bypass firewalls that prohibits or filter certain internet services.
In other words SSH Tunnels can be used to get around firewall filtering. If you never heard of Bypassing firewalls with port forwarding I recommend to check this article

We can use a ssh tunnel to securely transfer files between a FTP server and a client even though the FTP
protocol itself is not encrypted
, Tnnel traffic to an SQL server, Tunnel traffic from your Desktop PC to a Proxy
SSH tunnels can be used to tunnel outbound E-mail traffic back to your work PC to avoid having to change SMTP servers etc.

A very common SSH Tunnel use is by office workers who don't want their Custom Web browsing habbits and Web history be tracked,

For people who has to travel a lot tunneling SSH traffic from notebook to your home based SMTP is mandatory, otherwise your e-mail passwords, might end up captured by someone who manages the Free WI-FI used by you somewhere on the road or cafeteria. Another good case use of SSH Tunnel is to tunnel VNC traffic.

SSH Tunneling was historically very famous in days where IRC (Internet Relay Chat) and used to be common way (for people who want to prevent their real IP address from publicly visible) on the Internet.
SSH Tunneling is great way to maintain high communication security, however because it requires some basic technical knowledge most people neglect it or never heard of it, so I hope this article will give at least basic ideas to people new to tunneling.

How to create SSH Tunnels on Microsoft Windows

SSH Tunneling has been originally invented on UNIX / Linux platform, luckily nowdays it is easily possible to create SSH tunnels on almost any moden OS (including Mac OS X and M$ Windows).

Using Plink it is possible to create easily (batch) scriptable SSH tunnels on Windows

Generally here is example of basic plink use syntax:
 

plink.exe {remote-host} -P 22 -C -D 10080 -l username -pw password
 

To make a Tunnel with Plink to remote SSH Server (remote-ssh-server.com) listening for connections on port 22 on localhost 10022

"C:Program Files (x86)PuTTYplink.exe" -ssh remote-username@tunnel-to-ssh-server-host.com -pw PASSWORD123 -C -T -D 127.0.0.1:10022 -N

-D – option instructs Plink to make "Dynamic SOCKS-based port forwarding"
Dynamic Socks-based port forwarding will make your ssh client a simple socks server which listen to port 10022 for incoming connections.
-Ttells plink not to open interactive shell on remote server (disable pty allocation)
-Cenables traffic compression

Using the -D option you can prepare tunnel to traffic all your outgoing server traffic via the SSH tunneling host.

E. g. -D allows you to use the remote ssh server as a proxy server with no need for remote server to run anything except SSH service and have Internet connection.

-N – stands for "don't start shell/command SSH-2 only"

On success of SSH Tunnel establishing you will get a pop-up window similar to below screenshot (note that some of the messages in cmd line are from a batch script – if no batch script is used you wil only get a window with "Using Username":

plink_establishing_ssh_tunnel-on-microsoft-windows-screenshot

Note: That if you're about to be using Plink command frequently add it to your Windows SYSTEM PATH (check my previous article – how to add Putty to System PATH) to enable it invokable without writting the full command path location.

Once this Window is open if you still have doubt, where the tunnel is established, the quickest way to test whether tunnel is working is to open telnet to localhost port 10022.

In Windows command prompt type:
 

C:> telnet localhost 10022

 

You will get a remote server SSH version printed, like on below shot:

windows-testing-whether-ssh-tunnel-is-working-with-telnet-screenshot-black-screen

Now lets say now you would like to access the internet via just created SSH tunnel.
You can do it by simply setting Socks Proxy in your Firefox / Opera (or whatever browser you use).

Whether you want to run the SSH Tunnel permanent and periodically check whether tunnel is allve and respawn it in case plink quit, you can use a very simple batch script like one below:
 

@ECHO OFF
tasklist /FI "IMAGENAME eq plink.exe" 2>NUL | find /I /N "plink.exe">NUL

if "%ERRORLEVEL%"=="0"  ( echo Programm is running exiting
exit
)
echo "Not Running"
start /b "C:Program Files (x86)PuTTYplink.exe" -ssh remote-username@tunnel-to-ssh-server-host.com -pw Password123 -T -C -D 127.0.0.1:10022 -N"


If you're a Linux guy and you don't want to mess around with Plink but still want to create your SSH tunnels following SSH client UNIX command line syntax, setup the Cygwin port of OpenSSH for Windows.

Or if you're unsure whether remote server is always reachable (you're moving with your notebook from (Country) network to network), you can use also portqry – windows port scanner to check whether remote ssh server port is opened:
 

@ECHO OFF
tasklist /FI "IMAGENAME eq plink.exe" 2>NUL | find /I /N "plink.exe">NUL
if "%ERRORLEVEL%"=="0"  ( echo Programm is running exiting
pause
exit
)
portqry -n tunnel-to-ssh-server-host.com -p tcp -e 22
if %ERRORLEVEL%==0 echo tunnel-to-ssh-server-host.com Reachable Connecting ..
if NOT %ERRORLEVEL%==0 (echo tunnel-to-ssh-server-host.com.com Unreachable
exit)
echo "Not Running. Starting"
"C:Program Files (x86)PuTTYplink.exe" -ssh remote-username@tunnel-to-ssh-server-host.com -pw Password123 -T -C -D 127.0.0.1:10080 –N


Here is another sample use of Plink to create tunnel via SSH host tunnel-to-ssh-server-host.com  to listen for connections on localhost port 1234 and forward all incoming traffic to squid proxy server on 192.168.1.5 on port 3128.
 

C:Usersgeorgi>plink.exe -v -x -a -T -C -noagent -ssh -L 127.0.0.1:1234:192.168.1.5:3128 remote-username@tunnel-to-ssh-server-host.com

Setting then 127.0.0.1:1234 in  Firefox Proxy browser settings will make all traffic from your browser to flow securily to your own proxy server (letting you hide your "custom traffic" from company Web sniffers (Proxying) and Web filtering.
A remark to make here is 192.168.1.5 is (a internal server with Squid Proxy, configured to pass traffic to the internet ).

Here is a way to make tunnel between your remote SSH server and local Desktop PC to make Google queries without being logged by your ISP or company where you're working:

plink.exe "-v -x -a -T -C -noagent -ssh -L 127.
0.0.1:1234:95.158.130.242:443 -pw SecretPassword123 -l hipo pc-freak.net"

95.158.130.242 – is one of the IPs www.google.com resolves to
pc-freak.net – is name of my home router.

To make tunneled queries to Google then open in browser https://127.0.0.1:1234

using_localhost_ssh_tunnel_to_anonymously-search-in-google-plink-tunnel-for-security

 

If you're a lazy windows user and don't want to bother to make your tunnels from command line and you want need nice gui way to manage multiple tunnels – check out putty-tunnel-manager.

I myself am not a SSH Tunnel expert and thus Iwould be happy to learn further from people for some interesting cases and custom ways to do  SSH Tunnels. If you know of better advantageous ways to tunnel traffic, please share.

Happy tunneling ! 🙂

Share this on