Fake Anderson Powerpole

I recently replaced a set of winch connectors on a Land Cruiser rescue-vehicle. The vehicle is about 10 years old, but I do not know the age of the connectors, they may have been replaced previously. They have been fitted to the vehicle since it was new, and were mounted underneath the bumper front and back. Local regulations makes permanent winch mounts difficult on road legal vehicles, so the winch was stored in the trunk. The connector looks like and mates with the original Anderson Powerpole SB 175 connectors, but closer inspection revealed that the connector was un-keyed. The original connectors are keyed such that you can only mate connectors of the same color, with the exception of black connectors that are un-keyed. This one was red, so it should be keyed. The Anderson site contains detailed drawings of the original connector. If you are unfamiliar with the SB-series of connectors and want to learn more I recommend taking a look at the catalog and mounting instructions.

The connectors show signs of heavy corrosion, and the green death is prevalent on the contacts. They were fitted with a plastic end cap, but it only offers limited protection against the environment. There is also a power shut-off switch for these connectors made to limit the corrosion caused by electrical current when the winch is not in use, but it has probably been left in the “ON” position for some months. Interestingly, the actual wire contact lugs appear to be original Powerpole contacts. It is a bit difficult to make out on the photo as the contact is upside down, but the “A” stamp was clearly visible underneath the green death. I guess the reason is that the original contacts are relatively cheap if you buy bulk, whereas the housings are expensive.

 

clip_image001

 

clip_image001

 

The housing on the left was mounted underneath the rear bumper, the one on the right was mounted on the winch and was mostly kept in the trunk. The housing broke apart when I tried to remove the cable. There is a spring in the part that is broken off that is supposed to hold the contact in place. This spring was completely corroded, thus proper removal of the contact was impossible. As you can see there is also a bolt and captive nut stuck to the connector housing. There are two apparent reasons for this; the bolt is corroded (even though it is actually stainless) and the steel panel has collision damage. Someone had backed into something, buckling the steel panel under the car where the captive nuts were mounted. I was unable to un-screw the captive nut, it just rotated in its’ hole, so I had to employ the “Clarkson-method”, using hammers and pry-bars.

 

clip_image001[5]

I was unable to verify if the captive nuts were original. The Land Cruiser has a lot of captive nuts installed at the factory, but these may have been installed with the winch kit.

Script to migrate VMs back to their preferred node

Situation

You have a set of VMs that are not running at their preferred node due to maintenance or some kind of outage triggering unscheduled migration. You have set one (and just one) preferred host for all you VMs. You have done this because you are want to balance your VMs manually to guarantee a certain minimum of performance to each VM. By the way, automatic load balancing cannot do that, there will be a lag between a usage spike and load balancing if load balancing is required. But I digress. The point is, the VMs are not running where they should, and you have not enabled automatic failback because you are afraid of node flapping or other inconveniences that could create problems. Hopefully though, you have some kind of monitoring in place to tell you that the VMs are restless and you need to fix the problem and subsequently corral them into their designated hosts. Oh, and you are using Virtual Machine Manager. You could do this on the individual cluster level as well, but that would be another script for another day.

If you understand the scenario above and self-identify with at least parts of it, this script is for you. If not, this script could cause all kinds of trouble.

Script Notes

  • You can skip “Connect to VMM Server” and “Add VMM Cmdlets” if you are running this script from the SCVMM PowerShell window.
  • The MoveVMs variable can be set to $false to get a list. this could be a smart choice for your first run.
  • The script ignores VMs that are not clustered and VMs that does not have a preferred server set.
  • I do not know what will happen if you have more than one preferred server set.

The script

#######################################################################################################################
#   _____     __     ______     ______     __  __     ______     ______     _____     ______     ______     ______    #
#  /\  __-.  /\ \   /\___  \   /\___  \   /\ \_\ \   /\  == \   /\  __ \   /\  __-.  /\  ___\   /\  ___\   /\  == \   #
#  \ \ \/\ \ \ \ \  \/_/  /__  \/_/  /__  \ \____ \  \ \  __<   \ \  __ \  \ \ \/\ \ \ \ \__ \  \ \  __\   \ \  __<   #
#   \ \____-  \ \_\   /\_____\   /\_____\  \/\_____\  \ \_____\  \ \_\ \_\  \ \____-  \ \_____\  \ \_____\  \ \_\ \_\ #
#    \/____/   \/_/   \/_____/   \/_____/   \/_____/   \/_____/   \/_/\/_/   \/____/   \/_____/   \/_____/   \/_/ /_/ #
#                                                                                                                     #
#                                                   http://lokna.no                                                   #
#---------------------------------------------------------------------------------------------------------------------#
#                                          -----=== Elevation required ===----                                        #
#---------------------------------------------------------------------------------------------------------------------#
# Purpose: List VMs that are not running at their preferred host, and migrate them to the correct host.               #
#                                                                                                                     #
#=====================================================================================================================#
# Notes:                                                                                                              #
# There is an option to disable VM migration. If migration is disabled, a list is returned of VMs that are running at #
# the wrong host.                                                                                                     #
#                                                                                                                     #
#######################################################################################################################
 
 
 
$CaptureTime = (Get-Date -Format "yyyy-MM-dd HH:mm:ss")
Write-Host "-----$CaptureTime-----`n"
# Add the VMM cmdlets to the powershell
Import-Module -Name "virtualmachinemanager"
 
# Connect to the VMM server
Get-VMMServer –ComputerName VMM.Server.Name|Select-Object Name
 
#Options
$HostGroup = "All Hosts\SQLMGMT\*" #End this with a star. You can go down to an individual VM. All Hosts\Hostgroup\VM.
$MoveVMs = $true #If you set this to true, we will try to migrate VMS to their preferred host.
#List VMS in the host group
$VMs = Get-SCVirtualMachine | where { $_.IsHighlyAvailable -eq $true -and $_.ClusterPreferredOwner -ne $null -and $_.HostGroupPath -like $HostGroup }
 
# Process
Foreach ($VM in $VMs) 
{
    # Get the Preferred Owner and the Current Owner
    $Preferred = Get-SCVirtualMachine $VM.Name | Select-Object -ExpandProperty clusterpreferredowner
    $Current = $VM.HostName
 
 
    # List discrepancies
    If ($Preferred -ne $Current) 
    {
        Write-Host "VM $VM should be running at $Preferred but is running at $Current." -ForegroundColor Yellow
        If ($MoveVMs -eq $true)
        {
            $NewHost = Get-SCVMHost -ComputerName $Preferred.Name
            Write-Host "We are trying to move $VM from  $Current to $NewHost." -ForegroundColor Green
            Move-SCVirtualMachine -VM $VM -VMHost $NewHost|Select-Object ComputerNameString, HostName
        } 
    }
}

Hyper-V VM with VirtualFC fails to start

Problem

This is just a quick note to remember the solution and EventIDs.

The VM fails to start complaining about failed resources or resource not available in Failover Cluster manager. Analysis of the event log reveals messages related to VirtualFC:

  • EventID 32110 from Hyper-V-SynthFC: ‘VMName’: NPIV virtual port operation on virtual port (WWN) failed with an error: The world wide port name already exists on the fabric. (Virtual machine ID ID)
  • EventID 32265 from Hyper-V-SynthFC: ‘VMName’: Virtual port (WWN) creation failed with a NPIV error(Virtual machine ID ID).
  • EventID 32100 from Hyper-V-VMMS: ‘VMNAME’: NPIV virtual port operation on virtual port (WWN) failed with an unknown error. (Virtual machine ID ID)
  • EventID 1205 from Microsoft-Windows-FailoverClustering: The Cluster service failed to bring clustered role ‘SCVMM VM Name Resources’ completely online or offline. One or more resources may be in a failed state. This may impact the availability of the clustered role.

Analysis

The events point in the direction of Virtual Fibre Channel or Fibre Channel issues. After a while we realised that one of the nodes in the cluster did not release the WWN when a VM migrated away from it. Further analysis revealed that the FC driver versions were different.

SNAGHTML66058b10

SNAGHTML660875d4

Solution

  • Make sure all cluster nodes are running the exact same driver and firmware for the SAN and network adapters. This is crucial for failovers to function smoothly.
  • To “release” the stuck WWNs you have to reboot the offending node. To figure out which node is holding the WWN you have to consult the FC Switch logs. Or you could just do a rolling restart and restart all nodes until it starts working.
  • I have successfully worked around the problem by removing and re-adding the virtual FC adapters n the VM that is not working. I do not know why this resolved the problem.
  • Another workaround would be to change the WWN on the virtual FC adapters. You would of course have to make this change at the SAN side as well.

Is your LAPS working as it should?

Intro

So, you have implemented LAPS, and you are wondering whether or not it is working as it should? Or at least, you should wonder about that. You see, LAPS is a solution with quite a few “moving parts”, and all of them have to work for your local administrator passwords to be randomized and rotated automatically. You need a Group Policy Client Side Extension on each and every Server and Workstation (client), you need a GPO using said extension, and you need to extend the schema and set AD permissions. If any of these are not working properly somewhere, LAPS will not work properly. The most usual problems are:

  • The GPO CSE is not deployed to some clients.
  • The GPO is not linked in all OUs where you have clients.

 Detection

We can easily check if LAPS is working for a specific client be reading the contents of the AD attributes associated with LAPS. We need access to read these properties, so all automated and manual testes mentioned henceforth has to be run by an account with permissions to read the properties. The LAPS operations guide details how you should configure the permissions. That being said, you should of course also test the permissions to make sure that only privileged users are able to read said properties. The properties are called:

  • ms-Mcs-AdmPwd stores the password as clear text.
  • ms-Mcs-AdmPwdExpirationTime stores the point in time for the next password change. The GPO checks this value when it is applied and resets the password if the time has passed.

We can use both of these to test if LAPS has been applied to a specific computer object at least once. If you do a manual test by using the Attribute Editor in AD Users and Computers you will see both. I have written PowerShell commands to automate the process based on the value of the ms-Mcs-AdmPwdExpirationTime attribute.

List computers without LAPS

This lists all computer objects without a LAPS expiry set. Virtual cluster computer objects are excluded. The results are exported to the file C:\TEMP\NoLaps.csv.

get-adcomputer -Properties Name, operatingSystem, Description, ms-Mcs-AdmPwdExpirationTime `
-LDAPFilter "(&(!ms-Mcs-AdmPwdExpirationTime=*)(operatingSystem=Windows*)(!Description=ClusterAwareUpdate*)(!Description=Failover cluster virtual network name account))"|`
Select Name, operatingSystem, Description, ms-Mcs-AdmPwdExpirationTime| Sort-Object Name | export-csv C:\Temp\NoLaps.csv -Delimiter ";" -NoTypeInformation

List computers with expired LAPS

Lists all computer objects where LAPS has been applied at least once, where the expiration time has passed. These are usually computers that are not powered on, maybe removed but not properly deleted from the AD. The results are exported to the file C:\TEMP\ExpiredLaps.csv.

$now = Get-Date
get-adcomputer -Properties Name, operatingSystem, Description, ms-Mcs-AdmPwdExpirationTime `
-LDAPFilter "(&(ms-Mcs-AdmPwdExpirationTime=*)(operatingSystem=Windows*)(!Description=ClusterAwareUpdate*)(!Description=Failover cluster virtual network name account))"|`
Select Name, operatingSystem, Description, @{N='ExpiryTime'; E={[DateTime]::FromFileTime($_."ms-Mcs-AdmPwdExpirationTime")}}| `
Where-Object ExpiryTime -lt $now| Sort-Object ExpiryTime| export-csv C:\Temp\ExpiredLAPS.csv -Delimiter ";" -NoTypeInformation

Get LAPS Expiration date for one or more computer(s)

This command lists the expiration time for one or more computers based on an LDAP filter. The sample filter (Name=Badger*) will list all computers whose name starts with Badger. Computers where the expiration time is not set are filtered out. For more information about the LDAP filter syntax se this link:  https://social.technet.microsoft.com/wiki/contents/articles/5392.active-directory-ldap-syntax-filters.aspx

$now = Get-Date
 get-adcomputer -Properties Name, operatingSystem, Description, ms-Mcs-AdmPwdExpirationTime `
-LDAPFilter "(&(ms-Mcs-AdmPwdExpirationTime=*)(Name=Badger*))"|`
Select Name, operatingSystem, Description, @{N='ExpiryTime'; E={[DateTime]::FromFileTime($_."ms-Mcs-AdmPwdExpirationTime")}}| Sort-Object Name

Get LAPS expiration date for one or more computers, excluding those with no expiry set

Similar to above, but includes computer objects where the expiration time is not set. Those return 01.01.1601 01.00.00 as ExpiryTime because of the conversion of 0 from FileTime to DateTime. To put it in another way, if the expiration time is reported as 01.01.1601 01.00.00 it has not been set.

$now = Get-Date
 get-adcomputer -Properties Name, operatingSystem, Description, ms-Mcs-AdmPwdExpirationTime `
-LDAPFilter "(&(!ms-Mcs-AdmPwdExpirationTime=*)(Name=Badger*))"|`
Select Name, operatingSystem, Description, @{N='ExpiryTime'; E={[DateTime]::FromFileTime($_."ms-Mcs-AdmPwdExpirationTime")}}| Sort-Object Name

WN602v2 / WNR612v2 and OpenWrt

clip_image001

During my Christmas holiday I came across a couple of discarded NetGear WN602v2 boxes. They are clearly marked with NetGear, but also seem to self-identify as a ViasatOnDemand wireless device. From what I can gather, they are used as a wireless network bridge, wirelessly connecting a satellite receiver to a wired ethernet network. Viasat is a Scandinavian television distributor.

clip_image002

Research

I was curious as to how the device was configured, so I set out on a quest searching for answers. NMAP confirmed my suspicions that the IP was set to 192.168.1.1. I tried connecting to the device both with a browser and a console application with no success. After wading through some misses I came across this forum post revealing that the WN602v2 was in fact a special edition of the WNR612 B: https://dd-wrt.com/phpBB2/viewtopic.php?p=871408. Further research brought me to the OpenWrt page for the WNR612 v2: http://wiki.openwrt.org/toh/netgear/wnr612v2. Here I found a pinout for a serial port jumper that should be located on the board. I pulled out my trusty screwdriver kit and located the T9 screws underneath the rubber feet.

T9 is a bit of an odd Torx size, as Torx bit sets usually go from T10 and upward. Thus, you either have a boatload of these laying around or you have never heard about them. They can usually be found in comprehensive mobile repair toolkits or in “specialty bit sets” at your local low cost hardware monstrosity.

Once into the box, the jumper was easy to locate. The leftmost pin appears to be labeled 4, but I suspect this is really part of CA114. Anyways, it will henceforth be known as pin four as that corresponds with the data from OpenWrt.

clip_image003

I connected my BusPirate, making sure to not connect the 3.3V line to decrease the risk of releasing the magic blue smoke. After putting the BusPirate in UART pass-through and rebooting the board I was greeted with a OpenWrt console. Version 7.09 Kamikaze to be exact.

clip_image004

I was still unable to log in though, but the bootloader supports overwriting the firmware using TFTP. First I tried loading the latest OpenWrt, v15 at the time of writing. That failed miserably. No matter what I did, the upload would fail after a few seconds. I tried multiple TFTP clients, different cables and whatnot, but to no avail. A Wireshark capture revealed that the TFTP did not receive any ACK-messages as soon as the transfer started. After some time I tried changing the network connection to half duplex 100MB on the client side. That did the trick and the firmware upload completed successfully. Or so I thought until the board restarted…

[ 7.280000] Kernel panic – not syncing: VFS: Unable to mount root fs on unknown-block(31,4)

[ 7.280000] —[ end Kernel panic – not syncing: VFS: Unable to mount root fs on unknown-block(31,4)

[ 82.490000] random: nonblocking pool is initialized

Not really what I call a success… But I persevered and decided to try the latest NetGear fw image for the WNR612v2. To my astonishment it worked perfectly, no problems at all. Some digging around in forums indicated that the load screen messages was probably caused by the firmware image being too big for the flash memory on the board. It was indicated that it was possible to reduce the size of the OpenWrt image by building your own custom image, but that was a bit far from my end goal. All I wanted was access to the box, and possibly to make it usable as a backup router. I knew it was able to run OpenWrt 7, so I decided to just try versions until I found one that worked, starting with v 14, the previous one. And it worked like a charm, ten minutes later I was greeted with the OpenWrt Barrier Breaker console.

clip_image005

I sourced a fitting ethernet RJ45 jack from the usual suspects in Hong Kong and soldered it to the board (shown fitted on the image above). From what I can figure out it should have been blue, but who cares? The next chapter details the process from start to finish without all the dead ends.

Flashing new firmware using TFTP and BusPirate

Just a side note: If you do not possess a BusPirate, do not get one for the sole purpose of flashing this router. The FTDI Friend + from AdaFruit is a lot cheaper and easier to work with. You may also need some jumper wires.

This procedure requires a basic understanding of electronics, ESD shielding and working with exposed circuit boards.

Equipment list

  • Windows Computer acting as TFTP client and terminal. You can probably use apples or penguins as an alternative in a pinch.
  • BusPirate to interface with the console. (FTDI Friend+ from AdaFruit also works, and probably other FTDI adapters as well.)
  • Software: RealTerm and TFTPD64.
  • Optional: USB Ethernet adapter.
  • Optional: RJ45 PCB connector, 20 x 15,5 x 14 mm, EAN 4894462487914 or similar.

Procedure

  • Dismantle the WN602v2. Four torx T9 screws are located under the rubber feet.
  • Remove the board from the case.
  • Wire the BusPirate to the console jumper. I did not connect the 3.3V line. Black wire to pin 2, Gray wire to pin 3 and Brown wire to pin 4.
  • Serial bus interface pinout, pin 4 closest to the heatsink:
  • 1. 3.3 V (Not connected to BusPirate)
  • 2. TX
  • 3. RX
  • 4. GND
  • BusPirate probe color codes:
  • clip_image006
  • clip_image007

    • If you use an FTDI Friend, refer to its manual for pinout. Remember to connect RX to TX and vice versa.
    • Connect to one of the yellow ethernet ports from the TFTP client computer. Use a USB ethernet dongle to reduce the risk of destroying your computer if something fails.
    • Power on the board.
    • Check for magic blue smoke. If you see it, abort.
    • Set the ethernet connection to 100mbps, half duplex. If you forget this, you get timeout error messages as the router is in half duplex mode and unable to transmit ACK-messages.
    • Set the IP for the connection to 192.168.1.2/24 or anything other than 192.168.1.1 (the router IP).
    • Connect to the BusPirate terminal, use Ansi mode in RealTerm.
    • Enter UART mode
    • Connect to the router console at 115200 baud, default settings for the other parameters.
    • Start UART passthrough macro (1).
    • (If you use a FTDI Friend, just connect it at 115200 baud)
    • Press a key when requested to interrupt boot and enter the bootloader
    • Execute “protect off all” to remove any write protection on the firmware memory.
    • Execute “fsload” to enter firmware recovery mode.
    • Send the firmware image using TFTP client mode on the computer.
    • clip_image008
    • Wait for the update process to complete. The router boots automatically once the process has finished.
    • Look at the boot messages to make sure the flash was successful.
    • Solder in a plug for the WAN ethernet port if you want to use it. I could only find a shielded version of a connector in the correct size, but the shielding was easily removed. The PCB does not support shielded connectors.

    • clip_image009
    • Dremel out a hole in the backplate for the new connector
    • Log in to OpenWrt and configure your new router.

    Securing Windows Active Directory

    This is a list of measures you can implement to increase your Windows AD Security. The list is in no way exhaustive, and some of the items overlap. Be aware that security recommendations change over time. This article was originally created 2018.01.22. If that is several years in the past when you read this, I cannot promise that all recommendations are up to date.

    LAPS – Local administrator password management

    Implementing LAPS ensures that all your domain-joined computers have a unique password that is changed periodically for the local administrator account. It operates as a GPO Client Side Extension, and thus requires you to install and register a DLL on each target computer. You can do this via GPO, in your VM image, or through any other software deployment solution you may use.

    On the management computers and/or the DC itself, you have to add management tools and GPO Editor templates. There is a graphical user interface and a PowerShell module. The PowerShell module also includes the commands necessary to extend the AD Schema for storing the passwords and their associated expiry date.

    See https://technet.microsoft.com/en-us/mt227395.aspx for details.

    Securing the built-in Administrator account

     

    The built in Administrator account in the domain should be secured. The ObjectSID of the domain admin account always ends in -500, and is thus easy to identify even if the name has been changed. The guidance used to be “Disable the Administrator account”, but it has been changed due to some recovery scenarios requiring an active Administrator-account. Specifically, the Administrator account is the only account able to log on when no global catalogs are online.

    See https://docs.microsoft.com/en-us/windows-server/identity/ad-ds/plan/security-best-practices/appendix-d–securing-built-in-administrator-accounts-in-active-directory for details and an implementation guide. Some highlights are shown below.

    Set the DOMAIN\Administrator account as sensitive and require smart card

     

    clip_image001

    Create a GPO to prevent Domain Admins from logging on to member servers or workstations

    I have gone a bit further than the guide here, adding Domain Admins and Guests for good measure. The “Local account and member of Administrators group” is related to denying local administrator accounts access to the computer from the network. More about this below.

    Make sure that this GPO does not apply to domain controllers, that is, do not link it at the domain level.

    clip_image002

     

    Block remote access for local accounts

    Add Guests, Local account and member of Administrators group, Domain Admins, Enterprise Admins and Schema Admins to the policy Computer Configuration\Windows Settings\Local Policies\User Rights Assignment\Deny Access to this computer from the network.

    clip_image003

    For details, see https://blogs.technet.microsoft.com/secguide/2014/09/02/blocking-remote-use-of-local-accounts/

    Disable weak ciphers for Windows Secure Channel

    You can build a GPO to limit the cipher suites used by the Windows Secure Channel API, and by extension IIS. Be aware that this does not in any way limit other usage of weak ciphers. For instance, a TomCat server running on the same computer may very well use RC4 even if you have removed it from the list of Windows secure channel ciphers.

    The GPO is located at Computer Configuration\Administrative Templates\Network\SSL Configuration Settings\Cipher Suites.

    When you enable this setting, you get a list of all the default ciphers as a long comma separated string. Which ciphers you get is dependent of the Windows version. The easiest way to edit this list is to copy the string into a text editor. You can change the order to change the priority and remove weak ciphers.

    clip_image004

    clip_image005

     

    Do not allow local users to run remote elevated sessions

    Do not apply this fix: https://support.microsoft.com/en-us/help/951016/description-of-user-account-control-and-remote-restrictions-in-windows

    That is, do not create the HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows\CurrentVersion\Policies\System \LocalAccountTokenFilter Policy value, and if it exists, make sure it is set to 0. We could of course create a GPO to enforce this setting.

    clip_image006

    For details, see https://www.harmj0y.net/blog/redteaming/pass-the-hash-is-dead-long-live-localaccounttokenfilterpolicy/

    Set a password policy and a lockout policy

     

    • Password length: 8 characters. Encourage users to create passwords with a random length between 8 and 20 characters. You want your users to have passwords that vary in length. If you set this limit to 14, chances are all passwords are exactly 14 characters long. This makes it a lot easier to crack them.
    • Complexity not required. If you require complexity, users tend to add numbers and capitals at the start and end of the password.
    • Password history: 10.
    • Maximum password age: 0, that is password never expires. To frequent password changes may lead to bad password diversity and predictable passwords. Leaked passwords are almost always exploited immediately, so there is no point in forcing a monthly password change. If you must, set the maximum age to one year. Urge users to choose new passwords that are completely different from the previous passwords. That is, do not use MypassWord1, MypasswOrd2 and so on.
    • Do not enable the reversible encryption option. Ever. Just don’t.
    • Lockout policy: Locked for 24 hours after five unsuccessful attempts.

    clip_image007

    clip_image008

    For background information, see:

    Enforce SMB Signing and disable SMB1

     

    Enforce signing

    You can enforce signing on both the server and client side. The server side is shown below. Be aware that some services require this setting to be disabled. If you have such services, create an overriding GPO for those servers only, leaving SMB signing on in the rest of the domain.

     

    clip_image009

    See https://technet.microsoft.com/en-us/library/cc731957(v=ws.11).aspx

    Disable SMB1

    You have to create some registry-GPO settings. Details are at the link below. Be aware that legacy clients like Windows XP will be dependent on SMBv1 on Domain Controllers to access the Sysvol share. The recommendation is still to disable SMBv1 everywhere.

     

    clip_image010

     

    See https://blogs.technet.microsoft.com/staysafe/2017/05/17/disable-smb-v1-in-managed-environments-with-ad-group-policy/ for details.

    Create new computer objects in a separate OU, not in the Computers container

     

    Thus you can delegate permissions to manage them, and you can apply GPOs to newly added computers. You do this with the rdircmp console application.

    • Log on to a domain controller.
    • Start an administrative CMD-shell
    • Execute rdircmp [FQDN of OU]

    clip_image011

    You can verify or check this setting using PowerShell:

    Get-ADDomain |Select-Object ComputersContainer.

    Limit the number of domain admins

    Domain admin accounts should only be used for domain administration tasks, and you should not have many of them. Do not use service accounts with domain admin access.

    A recommended number of domain admins is 5.

    Avoid explicit permissions, prefer group permissions

     

    All permissions in AD should preferably be given to groups, not individual users. This makes it a lot easier to manage permissions, and it is also easier to see what permissions a user has based on which groups he is a member of. That is, if you follow this principle. There will always be exceptions, but they should be few and far between.

    Limit the number of people with delegated access to AD

     

    AD administration task can be delegated. For instance, your service desk could be able to reset passwords and create users without full domain admin access. It is important to limit these delegations and keep tabs on them.

    Use dedicated domain controllers

     

    • Make sure that your domain has at least two domain controllers.
    • If they are virtual, they should not be on the same cluster. Preferably you should have at least one dedicated physical domain controller.
    • Do not install anything on you domain controllers, with the exception of backup agents, antivirus software, monitoring agents and software deployment agents.
    • Do not enable the Hypervisor role on your physical domain controllers to run other software in a VM.
    • Make sure you have a system state backup of your domain controllers.

    No trusts between domains

     

    Avoid using forests and trusts between domains. Trusted domains should be handled as a single security context (e.g. dev, test, production, management etc), and thus you only really need one domain for each security context.

    Enforce the Windows firewall

     

    Make sure that Windows Firewall is turned on. There are many ways to do this, e.g. SCCM or GPO.

    Install antivirus software on all servers and workstations

     

    And make sure that it is activated and up to date. SCCM enables you to monitor and manage the default Windows Defender antivirus. Most commercial AntiVirus software comes with some kind of centralized management and monitoring tool.

    Log out RDP sessions after 24 hours

     

    Remote desktop server sessions that are still active (idle or disconnected) after 24 hours should be logged out automatically. Really active sessions are left for 5 days.

    The GPO settings are located at:

    Computer Configuration\Administrative Templates\Windows Components\Remote Desktop Services\Remtoe Desktop Session Host\Session Time limits

    • Set time limit for diosconnected sessions: 1 day.
    • Set time limit for active but idle RDS sessions: 1 day.
    • Set time limit for active RDS sessions: 5 days.

    Group Managed Service accounts and Managed Service Accounts

    Enable the domain for group managed service accounts, and encourage its use on supported services.

    https://docs.microsoft.com/en-us/windows-server/security/group-managed-service-accounts/group-managed-service-accounts-overview

    https://blogs.msdn.microsoft.com/markweberblog/2016/05/25/group-managed-service-accounts-gmsa-and-sql-server-2016/

    Netgear TL-SG108

    This small SOHO-type switch was subject to an overvoltage on the network cable coming from an ADSL-modem. The modem survived, the switch and one of the attached computers did not. Apologies for the poor image quality, I did not bring my camera gear for this.

    clip_image001

    image

    In the close-up of the middle part we se that both R22, R24 and C75 are gone.R26 and R28 looks suspicious but may be OK. The Ethernet chips on the reverse side looks fine, and the self test (blinking lights) completed on four of the ports. There is a total of four resistor and capacitor sets, each connected to two ports. As shown in the close-up below, C76, R25 and R27 have also been relieved of their magic blue smoke. As the capacitor is shared between the ports in the set, none of them are working even if only two of the four ports have their resistors blown. It is likely that this is fixable with new resistors and capacitors, as their function is probably to shield the fancy Ethernet control chips from nasty voltages. But, as I do not carry the necessary parts and equipment it was easier to just replace the entire unit for the extortionist price of 30USD at the local brick and mortar store.

    image

    Removing a drive from a cluster group moves all cluster group resources to Available Storage

    Problem

    During routine maintenance on a SQL Server cluster we were planning to remove one of the clustered drives. We had previously replaced the SAN, and this disk was backed by an old storage unit that we wanted to decommission. So we made sure that there were no dependencies, right-clicked the drive in Failover Cluster Manager under the SQL Server role and pressed “Remove from SQL Server”. Promptly the drive vanished from view, together with all other cluster resources associated with the role…

    After a slightly panicky check to make sure that the SQL Server instance was still running (it was), we started to wonder about what was happening. Running Get-ClusterResource in PowerShell revealed that all our missing resources had been moved to the “Available Storage” resource group.

    image

    We did a failover to verify that the instance was still working, and it gladly failed over with the Available Storage group. There is a total of 4 instances of SQL Server on the sample cluster pictured above.

    Solution

    The usual warning: Performing this procedure may result in an outage. If you do not understand the commands, read up on them before you try.

    Move the resources back to the SQL Server resource group. If you move the SQL Server resource, that is the resource with the ResourceType SQL Server, all other dependent resources should follow. If your dependency settings are not configured correctly, you may have to move some of the resources independently.

    Command: Get-ClusterResource “SQL Server (instance)”|Move-ClusterResource –Group “SQL Server (instance)”

    Just replace Instance with the name of your SQL Server instance.

    Then, run Get-ClusterResource|Sort-Object OwnerGroup, ResourceType to verify that all you resources are associated with the correct resource group. The result should look something like this. As a minimum, you should have an IP address, a network name, SQL Server, SQL Server Agent and one ore more Physical disk drives.

    image

    Microsoft Update with PSWindowsUpdate 2.0

    Preface

    This is an update to my previous post about PSWindowsUpdate located here: https://lokna.no/?p=2132. The content is pretty much the same, but updated for PSWindowsUpdate 2.0.

    Most of my Windows servers are patched by WSUS, SCCM or a similar automated patch management solution at regular intervals. But not all. Some servers are just too important to be autopatched. This is a combination of SLA requirements making downtime difficult to schedule and the sheer impact of a botched patch run on backend servers. Thus, a more hands-on approach is needed. In W2012R2 and far back this was easily achieved by running the manual Windows Update application. I ran through the process in QA, let it simmer for a while and went on to repeat the process in production if no nefarious effects were found during testing. Some systems even have three or more staging levels. It is a very manual process, but it works, and as we are required to hand-hold the servers during the update anyway, it does not really cost anything. Then along came Windows Server 2016. Or Windows 10 I should really say, as the Update-module in W2016 is carbon copied from W10 without changes. It is even trying to convince me to install W10 Creators update on my servers…

    clip_image001

    In Windows Server 2016 the lazy bastards at Microsoft just could not be bothered to implement the functionality from W2012R2 WU. It is no longer possible to defer specific updates I do not want, such as the stupid Silverlight mess. If I want Microsoft update, then I have to take it all. And if I should become slightly insane and suddenly decide I want driver updates from WU, the only way to do that is to go through device manager and check every single device for updates. Or install WUMT, a shady custom WU client of unknown origin.

    I could of course use WSUS or SCCM to push just the updates I want, but then I have to magically imagine what updates each server wants and add them to an ever growing number of target groups. Every time I have a patch run. Now that is expensive. If I had enough of the “special needs” servers to justify the manpower-cost, I would have done so long ago. Thus, another solution was needed…

    PSWindowsUpdate to the rescue. PSWindUpdate is a Powershell module written by a user called MichalGajda enabling management of Windows Update through Powershell. You can find it here: https://www.powershellgallery.com/packages/PSWindowsUpdate/2.0.0.0. In this post I go through how to install the module and use it to run Microsoft Update in a way that resembles the functionality from W2012R2. You could tell the module to install a certain list of updates, but I found it easier to hide the unwanted updates. It also ensures that they are not added by mistake with the next round of patches.

    Getting started

    (See the following chapters for details.)

    • You should of course start by installing the module. This should be a one-time deal, unless a new version has been released since last time you used it. New versions of the module should of course be tested in QA like any other software.
    • Then, make sure that Microsoft Update is active.
    • Check for updates to get a list of available patches.
    • Hide any unwanted patches
    • Install the updates
    • Re-check for updates to make sure there are no “round-two” patches to install.

    Continue reading “Microsoft Update with PSWindowsUpdate 2.0”

    About the UserAccountControl attribute

    Intro

    When working with MIM you will sooner or later have to deal directly with the UserAccountControl Active Directory attribute. This attribute defines account options, and we use it most prevalently to enable and disable users, but there are a lot of other options as well. These options are stored in a binary value as bit flags, where each bit defines a specific function.

    Bit number 1 (or 2 if you are not used to zero-based numbering) defines whether or not an account is enabled. Bit number 9 defines an account as a normal account. Thus, a normal disabled account will have bits 1 and 9 set to one. As long as no other bits are set, the decimal value is 2^9 + 2^1 = 514 or (0010 0000 0010). If we enable the account, the value is 2^9 = 512 (0010 0000 0000).

    In MIM we are usually presented with decimal values. These are easier to read, but not necessarily easier to understand.

    Continue reading “About the UserAccountControl attribute”