SPF exists for a reason, and postfix makes it ridiculously easy

In the bronze age, messages came in, signed bob@lamuella.com, and you’d simply have to assume they were from the Almighty Bob. Today, email is still technically in the iron age, but at least we have SPF.

In a nutshell, when lamuella.com enables SPF for their domain, they can be sure that other people, which mail servers check those records, will not receive spam mail with forged senders @lamuella.com. This also means email that passes this sort of check can be rated as slightly less spammy in automated spam filters.

To further elaborate on that, here’s what SPF is, does, it not and does not:

SPF IS/DOES

  • Allow a host to verify that an email from an SPF-enabled domain was sent from a mail server that legitimately serves that domain.
  • Prevent spammers from sending email from your SPF-enabled domain.
  • Force your users to send their email through your (hopefully extra extra safe) server, thus forcing them to comply with your encryption routines and filters.
  • Allow you to easily discard lots of spam mail, as such junk is often sent from forged domain names that happen to exist and have valid SPF records.

SPF IS/DOES NOT

  • Allow your server to verify the sender address from non-SPF domains.
  • Prevent spammens from sending email from your SPF-enabled domain to a recipient server that doesn’t inspect SPF records.

That said, SPF is easy to enable on Postfix, and I strongly urge you to do so if you haven’t already.
I found this handy guide, for Ubuntu, but it works just as well on Debian, and I archived it here in case the link is dead at the time you read this.

pfSense as a Cisco AnyConnect VPN Client using OpenConnect

pfSense, as of 2016-03-01, does not support OpenConnect out of the box. However, it’s in the FreeBSD repository, and relatively easy to add:

# pkg
# pkg update -f
# pkg install openconnect
# rehash

You can now play around with the openconnect command and test your connection.

Next step: Autostart, and adding the tun interface to the pfSense GUI. The GUI will, by default, ignore any interface named “tun*”, while openconnect will refuse to work with any interface not named “tun*”. Brilliant. The easiest workaround for this special case seems to be renaming the VPN interface after creation.

I made a script that automates checking if the connection is up, and (re-)starting it if it is not.
Replace the options in the “settings” section with appropriate values for your setup, and you should be good to go.

The “test” field should be a command that returns 0 when the connection is up, and anything else when it’s broken. I used netcat’s port testing feature on the remote desktop port of a server I needed to be able to connect to, but you can just as easily use things like ping with a limited count or similar.

#!/bin/sh

# settings
user="vpnuser"
pass="P4ssw0rd"
host="vpn.server.here.com"
test="nc -v -w 10 -z 172.16.0.4 3389"
tmpif="tun69"
iface="ocvpnc1"
pidfile="/tmp/${iface}.pid"
script="/usr/local/sbin/vpnc-script"


# env
openconnect="/usr/local/sbin/openconnect"
ifconfig="/sbin/ifconfig"


# func
ifkill()
{
        $ifconfig "$1" down 2>/dev/null || :
        $ifconfig "$1" destroy 2>/dev/null || :
}


# check if we're already running
if [ -n "$test" ] && $test; then
        echo "Connection is already up"
        exit 0
fi


# clean up previous instance, if any
if [ -e "$pidfile" ]; then
        read pid <"$pidfile"
        echo "Killing previous pid: $pid"
        kill -TERM "$pid"
        rm "$pidfile"
fi
ifkill "$tmpif"
ifkill "$iface"


# open vpn connection
echo "$pass" |\
$openconnect \
        --background \
        --pid-file="$pidfile" \
        --interface="$tmpif" \
        --user="$user" \
        --passwd-on-stdin \
        --script="$script" \
        "$host"


# rename the interface
if [ "$iface" != "$tmpif" ]; then
        echo "Renaming $tmpif to $iface"
        $ifconfig "$tmpif" name "$iface"
fi

Next, use crontab -e and add an entry to run the script regularly.

*/5 * * * * /root/openconnect-vpn >/dev/null 2>&1

Again, replace the path and timing with your own preferred values.

With the connection established, you can now go ahead and add the interface in the “assignment” tab of the GUI and set up appropriate rules for it.

CAUTION: Adding an interface that’s not available at boot time to the GUI will cause pfSense to think something is wrong on subsequent reboots and ask you to configure interfaces. I am not currently aware of a workaround for this, other than to not add the interface, controlling rules directly from the script instead. Please use the workaround below to avoid this issue, and make sure to verify that it works before leaving a pfSense box at a remote site unattended.

Interface boot workaround

The following workaround was offered by “DJC” in the comments section:

  1. Install “Shellcmd” in PfSense WebConfigurator:
    System => Package Manager => Available Packages
    Find Shellcmd and INSTALL
  2. Navigate to Shellcmd:
    Services => Shellcmd
  3. Add the following item in Shellcmd:
    Command: /sbin/ifconfig tun create; /sbin/ifconfig tun0 name ocvpnc1
    Shellcmd Type: earlyshellcmd
    Description: Create tunnel interface for OVPNC1 at boot

Squid Compile with SSL support under Debian Jessie

Here’s how:

apt-get update

# install required dev packages
apt-get install -y openssl devscripts build-essential libssl-dev

# install debian squid3 source code
apt-get source -y squid3

# install all required dependeny packages
apt-get build-dep -y squid3

# reconfigure
cd squid3-3.4.8/
vi debian/rules

root@jessie:~/squid3-3.4.8# diff /tmp/rules debian/rules 
46c46,48
< --with-default-user=proxy
--- 
> 		--with-default-user=proxy \
> 		--enable-ssl \
> 		--with-open-ssl="/etc/ssl/openssl.cnf"

# build package
debuild -us -uc

# install
cd ..
apt-get install logrotate
dpkg -i *.deb
apt-get install -f

# test
root@jessie:~# squid3 -v
Squid Cache: Version 3.4.8
configure options:  '--build=x86_64-linux-gnu' '--prefix=/usr' '--includedir=${prefix}/include'
'--mandir=${prefix}/share/man' '--infodir=${prefix}/share/info' '--sysconfdir=/etc'
'--localstatedir=/var' '--libexecdir=${prefix}/lib/squid3' '--srcdir=.' '--disable-maintainer-mode'
'--disable-dependency-tracking' '--disable-silent-rules' '--datadir=/usr/share/squid3'
'--sysconfdir=/etc/squid3' '--mandir=/usr/share/man' '--enable-inline' '--disable-arch-native'
'--enable-async-io=8' '--enable-storeio=ufs,aufs,diskd,rock' '--enable-removal-policies=lru,heap'
'--enable-delay-pools' '--enable-cache-digests' '--enable-icap-client'
'--enable-follow-x-forwarded-for'
'--enable-auth-basic=DB,fake,getpwnam,LDAP,MSNT,MSNT-multi-domain,NCSA,NIS,PAM,POP3,RADIUS,SASL,SMB'
'--enable-auth-digest=file,LDAP' '--enable-auth-negotiate=kerberos,wrapper' '--enable-auth-ntlm=fake,smb_lm'
'--enable-external-acl-helpers=file_userip,kerberos_ldap_group,LDAP_group,session,SQL_session,unix_group,wbinfo_group'
'--enable-url-rewrite-helpers=fake' '--enable-eui' '--enable-esi' '--enable-icmp' '--enable-zph-qos'
'--enable-ecap' '--disable-translation' '--with-swapdir=/var/spool/squid3'
'--with-logdir=/var/log/squid3' '--with-pidfile=/var/run/squid3.pid' '--with-filedescriptors=65536'
'--with-large-files' '--with-default-user=proxy' '--enable-ssl'
'--with-open-ssl=/etc/ssl/openssl.cnf' '--enable-linux-netfilter' 'build_alias=x86_64-linux-gnu'
'CFLAGS=-g -O2 -fPIE -fstack-protector-strong -Wformat -Werror=format-security -Wall'
'LDFLAGS=-fPIE -pie -Wl,-z,relro -Wl,-z,now' 'CPPFLAGS=-D_FORTIFY_SOURCE=2' 'CXXFLAGS=-g -O2 -fPIE
-fstack-protector-strong -Wformat -Werror=format-security'

Shamelessly stolen from http://www.panticz.de/Squid-Compile-with-SSL-support-under-Debian-Jessie

Error 0x80041002 from PowerShell after upgrade to Windows 10

On a box I just upgraded, we run a script that’s supposed to add a new scheduled task. After the upgrade, attempting to create a new trigger would return this beauty:

PS C:\Users\User> New-ScheduledTaskTrigger -AtStartup
New-ScheduledTaskTrigger : Method "NewTriggerByStartup" not found
+ CategoryInfo          : ObjectNotFound: (PS_ScheduledTask:Root/Microsoft/...S_ScheduledTask) [New-ScheduledTaskTrigger], CimException
+ FullyQualifiedErrorId : HRESULT 0x80041002,New-ScheduledTaskTrigger

The solution was to run this magic batch file that re-registers a whole slew of DLL files with the system, and then reboot.

net stop winmgmt
c: 
cd %systemroot%\system32\wbem 
rd /S /Q repository
 
regsvr32 /s %systemroot%\system32\scecli.dll 
regsvr32 /s %systemroot%\system32\userenv.dll
 
mofcomp cimwin32.mof 
mofcomp cimwin32.mfl 
mofcomp rsop.mof 
mofcomp rsop.mfl 
for /f %%s in ('dir /b /s *.dll') do regsvr32 /s %%s 
for /f %%s in ('dir /b *.mof') do mofcomp %%s 
for /f %%s in ('dir /b *.mfl') do mofcomp %%s 
echo DONE

I found the batch file here, intended to fix another issue with the same error code.

A few group policies I use on Windows 10

Disable the lock screen, always showing a login prompt:
Group Policy Editor (gpedit.msc) => Computer Configuration => Administrative Templates => Control Panel => Personalization => Do not display the lock screen

Disallow the use of OneDrive:
Group Policy Editor (gpedit.msc) => Computer Configuration => Administrative Templates => Windows Components => OneDrive => Prevent the usage of OneDrive for file storage

Wiping an (Azure) SQL Server database, deleting all tables

Just for personal reference

USE [mydb]
GO

WHILE(EXISTS(SELECT 1 FROM INFORMATION_SCHEMA.TABLE_CONSTRAINTS WHERE CONSTRAINT_TYPE='FOREIGN KEY'))
BEGIN
  DECLARE @sql0 NVARCHAR(2000)
    SELECT TOP 1 @sql0=('ALTER TABLE [' + TABLE_SCHEMA + '].[' + TABLE_NAME
    + '] DROP CONSTRAINT [' + CONSTRAINT_NAME + ']')
    FROM information_schema.table_constraints
    WHERE CONSTRAINT_TYPE = 'FOREIGN KEY'

  EXEC (@sql0)
  PRINT @sql0
END

GO

DECLARE @sql1 NVARCHAR(2000)
DECLARE @sql2 NVARCHAR(2000)
WHILE(EXISTS(SELECT 1 FROM INFORMATION_SCHEMA.TABLES))
BEGIN
  DECLARE c1 CURSOR FOR
  SELECT
    ('DROP TABLE [' + TABLE_SCHEMA + '].[' + TABLE_NAME + ']') as sql1,
    ('DROP VIEW [' + TABLE_SCHEMA + '].[' + TABLE_NAME + ']') as sql2
    FROM INFORMATION_SCHEMA.TABLES
  OPEN c1
  FETCH NEXT FROM c1 INTO @sql1, @sql2
  WHILE @@FETCH_STATUS <> -1
  BEGIN
    EXEC (@sql1)
    PRINT @sql1
    EXEC (@sql2)
    PRINT @sql2
	FETCH NEXT FROM c1 INTO @sql1, @sql2
  END
  DEALLOCATE c1
END

GO

Mounting and using KVM raw image files on the host system

Boldly stolen from David Champion at the University of Chicago, and archived for personal reference. All credits to that page.

kpartx will scan a block device — e.g., your /dev/loop0 — and create device-mapped specials under /dev/mapper for the partitions. Then you can address individual partitions. (N.B.: When done, before you can losetup -d the greater block device, you have to delete the device-mapped specials!)

Here’s a bit of terminal action to illustrate:

# First map the loop device to your KVM disk image. Note that some older versions of losetup don't have the --show option, and will output the used loop device upon being given -v only
$ losetup -fv --show /kvm/watercooler.mwt2.org.img 
Loop device is /dev/loop0

# No partitions!    
$ ls /dev/loop0* /dev/mapper/loop0*
ls: /dev/mapper/loop0*: No such file or directory
/dev/loop0

# Use kpartx -a to discover and device-map the partitions.
$ kpartx -av /dev/loop0
add map loop0p1 : 0 204800 linear /dev/loop0 2048
add map loop0p2 : 0 16930816 linear /dev/loop0 206848
add map loop0p3 : 0 16416768 linear /dev/loop0 17137664

# Look, partitions.
$ ls /dev/loop0* /dev/mapper/loop0*
/dev/loop0  /dev/mapper/loop0p1  /dev/mapper/loop0p2  /dev/mapper/loop0p3

# Fiddle with /dev/loop0p1 et al.

# Try to un-loop the file.  You can't, because kpartx's partition devices are still mapped.
$ losetup -d /dev/loop0
ioctl: LOOP_CLR_FD: Device or resource busy

# Ask kpartx to unmap.
$ kpartx -dv /dev/loop0
del devmap : loop0p1
del devmap : loop0p2
del devmap : loop0p3

# Now you can un-loop.
$ losetup -d /dev/loop0

Hardening IIS and scanning for SSL-related vulnerabilities – IISCrypto

I stumbled upon this great tool to set all the relevant registry entries (WTF?) to allow only the cryptographic algorithms that are considered secure to be negotiated with the server.

I archived it here (CLI version), for my personal reference, but if you prefer (and you should, frankly) you can get it from the author’s site.

Also, make sure to scan your page using, for example, this tool when you’re done.

UPDATE (February 19, 2019): New version 3.0: GUI / CLI