I'm using CentOS 6.7 (Final) system, and when I try to run nc in listening mode it prints the following:
nc -l 1234
nc: Protocol not available
The port is not bound. I tried other port numbers too. This bug seem to have been reported already: https://access.redhat.com/solutions/1753753. Unfortunately it is not very detailed.
Package information:
Name : nc
Arch : x86_64
Version : 1.84
Release : 24.el6
Is there something else I need to try out?
I ran into the same issue. You can solve it this way:
# Removes the old package
yum erase nc
Manually downloads the working package from the Official Repository
wget http://vault.centos.org/6.6/os/x86_64/Packages/nc-1.84-22.el6.x86_64.rpm
Installs the package
rpm -iUv nc-1.84-22.el6.x86_64.rpm
Please note that the package is for x86_64 (64-bit). If you need i386 (32-bit), the correct one is:
wget http://vault.centos.org/6.6/os/i386/Packages/nc-1.84-22.el6.i686.rpm
This particular version of netcat has a bug. Until there's a fix out for it, the only thing you can do is to downgrade to a previous version - sudo yum remove nc-1.84-24.el6.x86_64; sudo yum install nc-1.84-22.el6.x86_64 should do the trick.
Replying the question:
a) -p is not supposed to be used when in listening mode. from nc manpage:
-l Used to specify that nc should listen for an incoming connection rather than initiate a connection to a remote host. It is an error to use this option in conjunction with the -p, -s, or -z options.
b) downgrade can be done in one step, yum downgrade command works with the url to the package:
$ rpm -q nc
nc-1.84-24.el6.x86_64
$ nc -l 12345 #Although the syntax is correct, the command fails
nc: Protocol not available
$ nc -l -p 12345 #attempt to run with incorrect syntax
usage: nc [-46DdhklnrStUuvzC] [-i interval] [-p source_port]
[-s source_ip_address] [-T ToS] [-w timeout] [-X proxy_version]
[-x proxy_address[:port]] [hostname] [port[s]]
$ sudo yum downgrade http://vault.centos.org/6.6/os/x86_64/Packages/nc-1.84-22.el6.x86_64.rpm #shortcut to downgrade
...
Setting up Downgrade Process
nc-1.84-22.el6.x86_64.rpm | 57 kB 00:00
Examining /var/tmp/yum-root-Iq4yc7/nc-1.84-22.el6.x86_64.rpm: nc-1.84-22.el6.x86_64
Resolving Dependencies
--> Running transaction check
---> Package nc.x86_64 0:1.84-22.el6 will be a downgrade
---> Package nc.x86_64 0:1.84-24.el6 will be erased
--> Finished Dependency Resolution
Dependencies Resolved
==============================================================================================================
Package Arch Version Repository Size
==============================================================================================================
Downgrading:
nc x86_64 1.84-22.el6 /nc-1.84-22.el6.x86_64 109 k
Transaction Summary
==============================================================================================================
Downgrade 1 Package(s)
Total size: 109 k
Is this ok [y/N]: y
...
Removed:
nc.x86_64 0:1.84-24.el6
Installed:
nc.x86_64 0:1.84-22.el6
Complete!
$ nc -l -p 12345 #attempt to run with incorrect syntax
usage: nc [-46DdhklnrStUuvzC] [-i interval] [-p source_port]
[-s source_ip_address] [-T ToS] [-w timeout] [-X proxy_version]
[-x proxy_address[:port]] [hostname] [port[s]]
$ nc -l 12345 # try to listen again
^C
$#nc successully opens a socket on 12345. had to stop it with ctrl+C
# nc -l -p 1234
You can listen on that port by explicitely using -p flag:
-p source_port
Specifies the source port nc should use, subject to privilege restrictions and availability.
The problem was along the lines of nc treating 1234 as the destination port (despite being run in listening mode), and trying to listen on port 0 (which produced that weird message). Explicitely passing -p works around it.
