Qsub: More than one host group - qsub

Is it possible to specify more than one host group in a call to qsub?
The specifications state that you can specify the host group with this syntax:
-q long.q##long.hgc (if long.hgc is a host group)
But I would like to know whether it is possible to specify more than one host group. There is nothing about that in the manual.

Just in case someone is curious about this, the answer was really straightforward:
-q long.q##long.hgc,long.q##long.hgd (if long.hgc and long.hgd are two host groups)

Related

Puppet create variable names using hiera

I want Puppet to create a different variable name depending on the hiera file associated with the environment. I want to do this because I want Puppet to use the ip address associated with a specific network interface. Ideally, the network interface will be in the hiera file. That way you could concatenate the ip_address variable name with the network interface defined in the hiera file, which would look something like.
::ipaddress_{$network_interface_from_hiera_file}
Is this possible?
Right now I have an the following, but I think there is a better implementation. If the network interfaces change I would have to add another case.
if $environment == 'production' {
$client_address = $::ipaddress_enp130s0f0
} else {
$client_address = $::ipaddress_eth2
}
It sounds like you're after an eval in Puppet, like you have in shell and Perl other languages, and as far as I know, there isn't one.
I would probably just use a custom fact that always returns the IP address I care about. Of course, then you need to solve the problem of how to get the custom facts out to your fleet.
Another solution might be to use Hiera's hierarchical lookup:
In hiera.yaml:
:hierarchy:
- %{::node_environment}
- common
In common.yaml:
---
myclass::client_address: "%{::ipaddress_eth2}"
In production.yaml:
---
myclass::client_address: "%{::ipaddress_enp130s0f0}"
Finally, be aware that you can look up values from within Hiera, see here. Possibly that could be helpful.

How does Mesos-DNS name tasks with slash ("nested")?

For example:
If Marathon is running a task named /cassandra, Mesos-DNS assigns it a DNS name - cassandra.marathon.mesos.
Now I have a task named /monit/promdash. How can I find its DNS name?
Already tried:
monit_promdash.marathon.mesos, promdash_monit.marathon.mesos (and with - instead of _), monit.marathon.mesos, promdash.marathon.mesos, ...)
There's a HTTP interface. Couldn't find how to list all DNS names either...
Thanks,
Marathon reverses the hierarchical names, concatenates them with - and this is the app name then, so in your case it would be promdash-monit.marathon.mesos. Try it out.
At the bottom of the Mesos-DNS naming documentation we provide some more details about how these FQHN are constructed and you can also check out a complete end-to-end example I've put together, using two levels of hierarchies.

Two values for same field with Jinja2 in Ansible

I'm trying to set up a template in ansible, for our tomcat servers, but we have two tomcat instances on each host, each of which needs a different value for certain variables, for instance:
Tomcat_1 needs a port set to 8105
Tomcat_2 needs a port set to 8205
Easy enough to do if it's only one value per node needed, but I'm having some trouble finding how to do this when you need multiple values per host in either the Ansible or jinja2 docs. Can anyone offer some assistance, or point me to an example?
What I'm thinking is something along the lines of if this filepath then this value, but I'm not sure how to make that happen with jinja2.
I would either use two roles or use the role syntax that allows you to pass in values...
- { role: tomecat, some_parameter: 3 }

TCP connection, bash only

I found this line in a script. While I globally understand what it does--opening a bidirectional TCP connection--, I need some explanations on the syntax. Here's the line:
exec 5<>"/dev/tcp/${SERVER}/${PORT}"
And my questions:
< and > are usually used to redirect IOs. What does it mean there? Is it usable in another context? How?
Why does it work, while /dev/tcp doesn't exists?
Why 5? Can it be another number? What are the values allowed?
Why is exec necessary? (given nothing is actually executed)
Thanks.
< and > are usually used to redirect IOs. What does it mean there? Is it usable in another context? How?
It's the same - input and output is redirected to fd 5.
Why does it work, while /dev/tcp doesn't exists?
It's a special file: If host is a valid hostname or Internet address, and port is an integer port number or service name, bash attempts to open a TCP connection to the corresponding socket.
Why 5? Can it be another number? What are the values allowed?
Yes, it can be any value, but you need to ensure you don't use an fd already in use.
Why is exec necessary? (given nothing is actually executed)
exec means the redirection happens in the current shell, not within a subshell.
I can only answer for the exec part:
exec without a command given may be used to change I/O redirections. <> in this case means open for read+write. 5 is the channel number (or file descriptor). This makes sense if other commands send their output / read their input from channel 5.
For "/dev/tcp/${SERVER}/${PORT}" I don't know if it's a feature of a specific Linux version or if it's a bash feature (I assume the latter).
-- EDIT: from the bash manual page: --
Bash handles several filenames specially when they are used
in redirections, as described in the following table:
/dev/fd/fd
If fd is a valid integer, file descriptor fd is
duplicated.
/dev/stdin
File descriptor 0 is duplicated.
/dev/stdout
File descriptor 1 is duplicated.
/dev/stderr
File descriptor 2 is duplicated.
/dev/tcp/host/port
If host is a valid hostname or Internet address,
and port is an integer port number or service
name, bash attempts to open a TCP connection to
the corresponding socket.
/dev/udp/host/port
If host is a valid hostname or Internet address,
and port is an integer port number or service
name, bash attempts to open a UDP connection to
the corresponding socket.

What is the easiest way using common linux tools to check if a bunch of ip addresses belongs to given network?

What is the easiest way using common linux tools to check if a bunch of ip addresses belongs to given network? I just need a number of how many of given addresses belongs to given subnet. Lets say network is 192.16.55.40/27 and addresses is 192.16.55.45, 192.16.55.115, 88.87.45.8, 192.16.55.37, 192.16.55.60 and 192.16.55.210..
I'm not sure whether you consider Ruby as a "common linux tool" but it has a nice module called IPAddr that has a method called include? for that.
require 'ipaddr'
net1 = IPAddr.new("192.168.2.0/24")
net2 = IPAddr.new("192.168.2.100")
net3 = IPAddr.new("192.168.3.0")
p net1.include?(net2) #=> true
p net1.include?(net3) #=> false
I needed this to, and decided to create a short script. I requires sed and bash. I'd call them both common linux tools.
Edit: Script too long to paste, apparently. You can find it here: http://folk.ntnu.no/olechrt/netaddr
$ cat ips
192.16.55.45
192.16.55.115
88.87.45.8
192.16.55.210.11
192.16.55.37
192.16.55.60
192.16.55.210
256.87.45.8
$ cat ips | netaddr 192.16.55.40/27
192.16.55.45
Warning: Input IP "192.16.55.210.11" is invalid.
192.16.55.37
192.16.55.60
Warning: Input IP "256.87.45.8" is invalid.
And finally, for the count you requested:
$ cat ips | netaddr 192.16.55.40/27 | wc -l
Warning: Input IP "192.16.55.210.11" is invalid.
Warning: Input IP "256.87.45.8" is invalid.
3

Resources