Hawkeye's questions

We know that Virtual Machines have a full OS with its own memory management installed with the associated overhead of virtual device drivers. In a virtual machine, valuable resources are emulated for the guest OS and hypervisor, which makes it possible to run many instances of one or more operating systems in parallel on a single machine (or host). Every guest OS runs as an individual entity from the host system.

We know that Docker Containers use runC (formerly known as libcontainer), which runs in the same operating system as its host. This allows it to share a lot of the host operating system resources. Also, it uses a layered filesystem (AuFS) and manages networking.

We know that Hyper-V implements isolation of virtual machines in terms of a partition. A partition is a logical unit of isolation, supported by the hypervisor, in which each guest operating system executes. There must be at least one parent partition in a hypervisor instance, running a supported version of Windows Server (2008 and later). The virtualization software runs in the parent partition and has direct access to the hardware devices. The parent partition creates child partitions which host the guest OSs. A parent partition creates child partitions using the hypercall API, which is the application programming interface exposed by Hyper-V.[9]

In this serverfault comment we read:

A Hyper-V container is something between a VM and a container, it runs it's own small copy of the chosen OS kernel, and as such behaves like a VM, and this is what enables networking. On the Windows side, it's basically like any other host running a Hyper-V VM with Linux in it.

My question is: Is a Hyper-V Container somewhere between a docker container and a virtual machine?

My understanding was that the primary limitation of running docker on other OSs was the Linux Network containers that made it possible. (Certainly for Macs).

Recently Microsoft announced a plan to provide a Linux kernel running natively packaged with Windows 10.

My question is: Can you run Docker natively on the new Windows 10 Linux kernel?

Note - this is not a duplicate of this question (which I also wrote) as that was about the Windows 10 bash userspace, and this is about the Windows 10 Linux kernel.

We have an encryption issue from a vendors mid-tier to an Oracle database. We can use encrypted database connections and have done for 90% of cases.

Our security team has advised us that since this connection goes over a link, everything needs to be secure.

One database client is a proprietary vendor mid-tier for whom we have asked to upgrade with encrypted SSL JDBC links, and they have advised us the timeline for this is two years. We need to have a security solution in a shorter timeframe.

Our security team advised us that stunnel might be a potential solution to this problem. This requires setting up a server, keys on the server and a client.

We already run an ssh server on the box that the Oracle database is running on, with keys for users setup.

It seems to me that stunnel duplicates the pattern of ssh.

1. server - stunnel server vs ssh server
2. server key setup - stunnel key setup vs openssh users setting their keys on the server
3. client - Using an ssh client vs using a stunnel client

Stunnel might be useful in cases where you're not already running an openssh server, or where you want a different key management system.

Now I might be missing something in my analysis. There might be a feature of stunnel that I need that I can't see.

My question is: Will ssh tunnelling be a simpler way to solve my link encryption issue rather than stunnel?

EDIT

• This link is already inside a secure network. It is not public. But it is financial services - where the threat model involves hard shell, soft core analysis. It's not good enough to have an insecure link inside a secure network - because you don't trust the inside of your network.

I've got the following file setting permissions on my machine:

/etc/sudoers.d/myuser

This has the following contents:

myuser ALL=(myuser) NOPASSWD: ALL

I'd like to add a second user to this list, something like:

myuser ALL=(myuser, myuser2) NOPASSWD: ALL

My question is: Can a line in a file in the /etc/sudoers.d directory have multiple users?

I'm trying to free up port 22 for a docker container (that I'll have to run in another port mapping).

I'm running

netstat -a | grep 22

One of the results I'm getting is:

kctl       0      0     22      9 com.apple.netsrc

My question is: What is kctl running on port 22 on a mac?

Many people are talking up the security benefits of Red Hat Atomic - being a read-only operating system - means substantially fewer vulnerabilities - particularly important in the cloud.

When I provision machine images with Packer - we install some software and then 'bake' an image - that we then use for other hosts.

To me it seems that with Red Hat Atomic -you must install some software than then 'bake' it in read-only mode.

In reading the documentation I can't see how to turn the key to lock it.

My question is: How do you 'bake' a Red Hat Atomic image in read-only mode?

I've got an AWS autoscaling configuration and a launch schedule that says to keep a minimum of 1 all day. I'd like to flush out this instance and get a new one every day. I'd like to maintain a minimum of 1 all day. Is this possible?

My question is: Can I force AWS to reinitialise my autoscaled instance daily maintaining a minimum of 1 instance?

I have a number of server applications running in Docker. The output is configured to go to an elk stack.

I've had a number of troubles with the elk stack and am considering going back to vanilla Docker logging. My concern is scalability.

My question is: How does Docker Daemon handle large log output?

My understanding was that the primary limitation of running docker on other OSs was the Linux Network containers that made it possible. (Certainly for Macs).

Recently Microsoft announced a beta of a Ubuntu linux user mode running natively on Windows 10. This can run binaries compiled in ELF format on Windows (unlike cygwin which requires a compilation.)

My question is: Can you run Docker natively on the new Windows 10 (Ubuntu) bash userspace?

I'm trying to install lynx on an AWS Redhat machine using the following:

[ec2-user@ip-XX-XX-XX-XX ~]$ sudo yum install lynx                                    
Loaded plugins: amazon-id, rhui-lb
rhui-REGION-client-config-server-7                       | 2.9 kB     00:00
rhui-REGION-rhel-server-releases                         | 3.7 kB     00:00
rhui-REGION-rhel-server-rh-common                        | 3.8 kB     00:00
rhui-REGION-rhel-server-releases/7Server/x86_64/primary_db |  14 MB   00:00
rhui-REGION-rhel-server-releases/7Server/x86_64/updateinfo | 639 kB   00:00
No package lynx available.
Error: Nothing to do

My question is: How to install lynx on an AWS redhat machine that can't find the package?

On a database, I can get a list of all the currently running processes, and the sql command that kicked them off.

I'd like to do a similar thing on a windows box.

I can get the list of processes, but not the command line that kicked them off.

My question is: Given a PID on Windows - how do I find the command line instruction that executed it?

Assumptions:

• Windows 7 and equivalent servers

Context - I'm trying to get a docker cluster running on MacOSX.

Part of the install process for getting docker running is to install pipework.

Pipework relies on /var/run/netns

My question is: Is there an equivalent of netns for Mac OS X?

You can see in the doco here that setting up each instance of Cassandra requires a new 'node'.

My question is - to run a cluster with multiple nodes of Cassandra on my Mac (Running Mavericks) - do I need to spin up a virtual machine for each - or can I run multiple instances (in different directories or with different config files for example)?

I'm trying to delete a user logged into my ubuntu system. Let's call the user bob.

I'm running

sudo userdel -r bob

The response I get is

userdel: user bob is currently logged in

So I run

ps -u bob

To find his bash session to kill it. But I'm not seeing it - I'm looking for something like:

13205 pts/0    00:00:00 bash

(Which is my session) but I don't see another one:

What command can I run to find out what this user is logged in as?

Edit:

who bob 

gives no result (to stdout at least)

last bob

gives a result three days ago - but running

sudo userdel -r bob

Still gives the same result.

Edit

who

only gives my login.

Edit

root@machine:~# sudo userdel -rf bob
userdel: user bob is currently logged in
userdel: warning: can't remove /var/mail/bob: No such file or directory
userdel: error removing directory /
root@machine:~# sudo userdel -rf bob
userdel: user 'bob' does not exist

This seems like it was not a correct type of user - is that correct?

I've got an existing software package that runs on Ubuntu that uses a chroot jail and so needs to be run as root. I want to exec this from a java web application running within Tomcat6 on Ubuntu, so presumably I have to run Tomcat6 as root instead of the Tomcat6 user.

How do I go about changing tomcat6 to run as root instead of the tomcat6 user?

Is it possible for a non-root user to run a chroot process on Ubuntu?

Has anyone seen this error on an XP desktop machine before?

'net' is not recognized as an internal or external command, operable program or batch file.

It's being used in a 'net start iisadmin' script

On Windows XP in an Active directory environment - what is the easiest way for me to query a user's email address from AD given their username on the command line.

(Assuming I know where it is kept normally in tree).

(I know about net user loginname /domain but I just want the email address element back.)

I want to implement a redundancy solution. I have two subnets all windows XP machines - each with their own DHCP server - joined by a bridge.

Currently - all machines get DHCP leases from the one DHCP server with the other disabled.

For when the bridge goes down - I want to split my DHCP scope among multiple servers. How can I do this on Windows XP client machines?

How can I set the priority of DHCP servers to get a lease from in windows (XP)?