Handling multiple overlapping VPN client networks with pfSense and a reflector VM

Following up on my recent post about making a minimal VPN routing virtual machine to isolate obnoxious VPN clients, another problem you’ll run into if you have the need to connect to a whole lot of client networks is that the networks of different companies tend to overlap, either with your own or with eachother. Most companies I deal with use the or the networks, as specified in RFC1918, or some combination thereof. seems to be really unpopular for some reason, even in small networks.

The problems arise when one client’s network fully or partially overlaps that of another client, or that of the company you’re at. If there’s a host in your network, and is also inside the client’s network, pfSense would have no idea where to route you. We could make a rule to route single computers to the remote site, but then they could never access the local machine with that IP. What we need is remapping, translation, dirty outbound NAT hacks and a bit of black magic.

A picture says a thousand words

This diagram attempts to show the address translation accomplished by pfSense and the reflector VM. The idea is to map the machines we’re trying to access at the remote site to an address range that is unused in the local network. Currently, we do this on an IP by IP basis, but it could easily be extended to translate entire address ranges if desired.

Image created with draw.io (XML).

Actually, that image didn’t explain a thing…

The idea here is that we have a client network that overlaps our own. We use an internally unused network, and connect to that instead. We exploit the fact that the routing table of an IPSec connection trumps the internal routing table in pfSense. A connection proceeds as such:

  1. The PC at attempts to connect to
  2. pfSense will outbound NAT this request and route it to the reflector VM at
    The reflector VM didn’t need an IP address this odd, but it has yet to collide with a client, so that’s nice.
  3. The request is now coming from the pfSense box, and arrives at the reflector as a connection attempt from pfSense to
  4. The reflector will also reverse NAT the request and route it back to pfSense as if it was making the request itself, but it will additionally translate the target IP from to
  5. pfSense receives the connection request from to, knows that is the local IP range for the VPN connection to Client1 and routes the connection over there.

Here are a few facts about this setup, to answer questions you may have at this point:

  • is set as the local IP range for the VPN connection, is set as the remote range
  • Computers in the local network, even those in other VLAN’s, such as the network, can thus not connect directly through the VPN, as their requests to IP’s would be routed as usual to the local 10.2 subnet. They all have to use to traverse the VPN’s
  • Computers at Client1 can’t possibly get back through this insane setup to reach local machines by IP
    However, they could make requests to the IP address we expose as the local part of our network in the NAT/BINAT configuration of IPSec, which is actually set up in a few cases, and the reflector can distribute those based on the port number and remote IP
  • The reflector VM can’t actually connect to internal machines with IP adresses overlapping the VPN networks, as its requests would be routed to clients
    ..but why would you want to use the reflector to connect to anything local?
  • If client networks overlap each other, you need more than one reflector

That’s completely insane!

I agree.

So how do we do it?

Glad you asked. Keep reading.

pfSense configuration

Firstly, we configure the VPN connection in pfSense as usual. Let’s assume it’s IPsec to make it easy. We set the local network to, and use NAT/BINAT translation to set the address to anything not in the client’s network, which will work just fine. The screenshot shows the IPSec Phase 2 configuration.

Next, we add the Reflector VM as a router on our internal network.

We also need to add a static route to send everything going to the network to the router we just created.

Then there’s the outbound NAT that translates requests going to the reflector as coming from pfSense. What we’re saying here is that connection requests going out of the TEST interface from any internal IP (I made this /8 just so I don’t have to touch it again) and heading to should be NAT’ed to the interface IP of pfSense.

Then we add the firewall rules. We need two of these. The first one allows the Reflector VM to connect to all networks, which includes all the remote clients. The second (not shown, but you’ll figure it out) allows the local machines to connect to the network.

This concludes the pfSense side of things.

Reflector VM

The reflector is a minimal install of Debian Linux with Shorewall installed. I’m familiar with both, so that makes it an obvious and time saving choice for me.

First, there’s the interfaces file. Just the single network interfaces in here. Not much to see, but the routeback option is needed to allow traffic to ingress and egress the same interface.

net eth0 detect dhcp,logmartians=1,nosmurfs,routefilter,tcpflags,routeback

The zones file is equally dull.

fw firewall
net ipv4

At the policy file, it gets a bit more interesting, as we allow any traffic to bounce off of us, going from net to net.

net net ACCEPT
all all REJECT info

As we want to NAT any traffic that bounces off us, we’ll need a masq file:

#                                                       GROUP        DEST
eth0 ,,

In shorewall.conf, I change a single line to ensure IP_FORWARDING is enabled


Then come the rules. This is where we do the mapping. These are mock entries to preserve company privacy. Note the single reverse connection too, which allows companies at the remote end to reach a local webserver.

#                   PORT  PORT(S)        DEST     LIMIT GROUP

# we want to SSH in here

# client1
DNAT any+ net: - - -
DNAT any+ net: - - -
DNAT any+ net: - - -

# client2
DNAT any+ net: - - -
DNAT any+ net: - - -
DNAT any+ net: - - -

# client3
DNAT any+ net: - - -
DNAT any+ net: - - -
DNAT any+ net: - - -

# reverse connections
DNAT any+ net: TCP 80

ACCEPT $FW net icmp
ACCEPT net $FW icmp

With our IP addresses re-mapped for the clients, everything should, amazingly, work. It’s not pretty. Far from it. But it does the job, it’s stable, and all I need to change to accommodate a new client is adding the map in the rules file after setting up the VPN.

The addresses in the range are completely arbitrary. I usually make the last 8 bits match the client’s machines, but you don’t have to do that.


Finally back to pfSense again, I also have some DNS host overrides so employees don’t have to remember the remapped, or the original, IP addresses. Note that if you’re dealing with HTTPS and want certificates to function correctly, the host.domain names need to match the certificates on the remote servers. These overrides are trivial to set up, but here’s an example anyway.

So what happens when two clients overlap with each other?

The above solution applies when clients partially or fully overlap your internal network. However, if they overlap with each other, pfSense would again not know where to route a given connection from the reflector. The solution then is another reflector VM. They’re very cheap to run anyway, so it’s not a big issue.

Final words

If you ever attempt to do this, you will no doubt have questions. Feel free to leave them here, and I’ll try to get back to you. Make sure you enter your e-mail correctly. It will not be published, but I need it if you want me to reply 🙂

Making a minimal VPN routing virtual machine to isolate obnoxious VPN clients

After it was requested multiple times on my post about using pfSense as a Cisco AnyConnect VPN Client, I finally found some time to document the setup I ended up with as a replacement, the purpose of which was to not clutter up the pfSense box. pfSense, after all, acts as the main company router and is integral to the operation of the business network. I don’t want to modify it beyond what some future software update would be able to handle.

The solution to this issue, and several others that have popped up since then, is a set of very small virtual machines. I run each of these, one for each non-pfSense supported client VPN solution (there are quite a few unusual ones out there), with 64 megs of RAM and a mere 3GB of harddrive space.

VM Setup

The virtual machines I use run Debian, as that’s what I’m most familiar with and when setting up things like these, I want them to consume as little of my time as possible. The OpenConnect VPN client is readily available in the Debian package repositories. No mucking about with the apt sources. I also use shorewall to do the routing and firewall rules for me. Thus:

  • Debian (minimal install)
  • OpenConnect (from repositories)
  • Shorewall
  • A cup of coffee


The script I use to keep it all going is pretty much copy-pasted from the post about using pfSense to do this, but here it is in its current form:


# settings
test="nc -v -w 10 -z 3389"

# env

# func
    $ifconfig "$1" down 2>/dev/null || :
    $ifconfig "$1" destroy 2>/dev/null || :

# check if we're already running
if [ -n "$test" ] && $test 2>/dev/null; then
    echo "Connection is already up"
    exit 0

# scream
echo "Connection is not up! Attempting restart..." >&2

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

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

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

I run the above script every minute using the following line in root’s crontab:

* * * * * /root/bin/cron/vpnscript >/dev/null


To do the routing and firewall, there are a few files in /etc/shorewall that are essential for this setup. The first is the interfaces file, in which I define the virtual machine’s regular network interface and the “tun69” interface created by the VPN connection.

net eth0 detect dhcp,logmartians=1,nosmurfs,routefilter,tcpflags,routeback
vpn tun69 detect dhcp,logmartians=1,nosmurfs,routefilter,tcpflags

Then there’s the zones file. This is pretty basic, but matches the interfaces.

fw firewall
net ipv4
vpn ipv4

For the policy file, we want to allow traffic to flow from the net zone to the vpn zone.

net vpn ACCEPT
all all REJECT info

The rules file allows SSH to control the VPN VM from the net zone, which is still in the company’s internal network, mind you.

#                   PORT  PORT(S)        DEST     LIMIT GROUP

We need a bit of magic to do outbound NAT, with the masq file. This makes the connections the policy file allows to flow from the net zone to the vpn zone masquerade as whatever IP address the VPN client got for its tun69 interface. Without this, the connections would get relayed using their original IP, which the VPN client network probably wouldn’t appreciate. This simply covers all RFC1918 IPv4 addresses, so it doesn’t rely in any way on the rest of the company’s network.

#                                                       GROUP        DEST
tun69 ,,

Finally, in shorewall.conf, I change a single line to ensure IP_FORWARDING is enabled


pfSense setup

Now we have a VPN routing machine, which in my case is situated on a dedicated VLAN to isolate it from the rest of the company network. You don’t have to do this, but if you don’t, anyone on the internal network can configure a route through the VPN machine and access whatever is on the other end.

In pfSense, I add a static IP for the OpenConnect client, then add a Gateway in System=>Routing=>Gateways. The interface should be the one the OpenConnect client is at, the IP should be the static IP of the client. Then, on the “Static Routes” page, add a route to the network at the end of the VPN and set the gateway to be the one you just created with the OpenConnect client IP. Now it’s up to your rules which machines you want to allow.

Any questions?

If you have any questions about this setup, let me know 🙂

BashDB, a single dynamic database table for bash scripts

I write shell scripts. A lot of them. I automate everything with scripts. Many of these scripts need to have some notion of a state. The state needs to be persistent. How do we do that? And could we draw some retro flashback line character table borders while we’re at it?

The problem

If I have some sort of a watchdog running every 5 minutes, I may not want to get an email about it until it has failed 12 times, or continuously for 60 minutes. I could run my script continuously and just sleep, or I could store the “fail count” in some file and read it back in on the next iteration. I often opt for the latter, and I used to find myself re-implementing pretty much the same tab separated, line separated, or null separated storage algorithms over and over.

Then come the lists. Say I have a script that’s supposed to verify that a couple of URL’s are online. Where do I store those? One per line in a file? I name it script.conf and put it somewhere? That’s fine, but what if I want to also store the last time each URL failed, so I can include more information in a report? Tab separated “columns”? A different file for the dates? Should I use a simple database, such as SQLite? That seems like a bit of an overkill, and it might not be installed where the script is supposed to be running.


Enter BashDB, written in the car going to and from a weekend trip at the cottage. It stores a single table, with a simple Key=>Value structure, or named columns. It dynamically adds new columns as requested, supports list values in any column, supports binary data, and it’s all written in Bash with very few external dependencies.

It can be downloaded –> HERE
View the manual –> HERE  (or run db_help after sourcing the library)


To use bashdb, simply source it in the top of your script and use the provided db_* functions:

db_help The equivalent of running each function, in turn, with the -? option. Shows the help 🙂
db_columns Lists the columns currently in the table
db_copy_row Copies a row with a given key to another row, optionally overwriting it if it exists
db_delete Deletes a row with a given key
db_delete_column Deletes one column of data from all rows
db_dump Provides a somewhat pretty display of the data currently in the table
db_get Gets a single value from the table
db_has_key Returns whether or not a given key exists
db_has_column Returns whether or not a given column exists
db_keys Lists the keys in currently in the table
db_rename Changes the key of a row
db_rename_column Renames a column
db_search Lists keys or dumps data matching a search regex in a given column
db_selftest Runs a unit test of most features
db_set Sets a value on a key
db_testdb Generates a small test database
db_trim Removes any empty columns


To try out the functionality, one can also use the script from the console, as such:

$ source ./bashdb
$ ls test
ls: cannot access test: No such file or directory
$ # we don't have a "test" yet :)

Adding a key and a value:

$ db_set
Error: db_set: Missing parameter: -f <file>
$ db_set -f test
Error: db_set: Missing parameter: -k <key>
$ db_set -f test -k Key1 -v Value1

Adding some more keys, and changing the value for key1:

$ db_set -f test -k "Norwegian Characters" -v "Ææ" -v "Øø" -v "Åå"
$ db_set -f test -k key1 -v "Value2"
$ db_set -f test -k key9 -v "Day[9]"

The “Norwegian Characters” is an example of storing a list. Lists can be read back out, either “human readable” -h, or null byte (\0) separated:

$ db_get -f test -k "norwegian characters"
ÆæØøÅå$ db_get -f test -k "norwegian characters" -h
$ db_get -f test -k "norwegian characters" -h | while read -r c; do echo "--> $c <--"; done
--> Ææ <--
--> Øø <--
--> Åå <--

We can unset values by setting a value to an empty string, “”, or by simply not including the -v switch at all, when calling db_set. A row will be removed entirely if all its values are unset.

Lists (which are actually any value at all), support the -m switch, followed by “add” or “remove” to modify the list:

$ db_set -f test -k key1
$ db_set -f test -k "norwegian characters" -v "Øø" -v "Åå" -m remove
$ db_set -f test -k key9 -m add -v "Day[10]"

Without specifying -c, a default unnamed column is used for the data. However, columns can be added dynamically to include more information about each key:

$ db_set -f test -k key9 -c "Additional" -v "Test Value"
$ db_set -f test -k key9 -c "Additional 2" -v "More stuff" -v "Here"
$ db_set -f test -k new -c "Additional 2" -v "Not all columns need to be filled"

Column data can be requested by using -c on db_get. Also showing an example of -d, giving a default value if none is set, as for key=>new, column=>additional:

$ db_get -f test -k key9 -c "additional 2" -h
More stuff
$ db_get -f test -k new -c "additional" -h -d "This is the default"
This is the default

Taking a short break to run the unit tests:

We can easily delete a row, or an entire column:

$ db_delete -f test -k new
$ db_delete_column -f test -c additional

Listing keys and columns, also with the -h option, since we’re showing it here.
The default column shows as an empty line.

$ db_keys -f test -h
norwegian characters
$ db_columns -f test -h

additional 2

We can also store values from stdin, for instance a web server response. This even works with binaries, though I wouldn’t recommend trying db_dump on tables containing those. Not that it breaks, it just looks like crap.

$ curl "https://google.com/" | db_set -f test -k "Response from google.com" -i
 % Total % Received % Xferd Average Speed Time Time Time Current
 Dload Upload Total Spent Left Speed
100 269 100 269 0 0 5458 0 --:--:-- --:--:-- --:--:-- 5489
$ db_get -f test -k "response from google.com"
<HTML><HEAD><meta http-equiv="content-type" content="text/html;charset=utf-8">
<H1>302 Moved</H1>
The document has moved
<A HREF="https://www.google.no/?gfe_rd=cr&amp;dcr=0&amp;ei=ej7bWYqlOuLk8AfSmqqoDg">here</A>.

When dumping a table for display, values are truncated. Please note that if you’re storing binary data, the displayed length by db_dump will only include the printable characters, and will not accurately reflect the length of the binary that would be retrieved by db_get.

Searching is done using a regex on a given column, optionally running db_dump on the results instead of listing the keys:

If you decide to use this for your own projects, and you’re having issues, please read the output of db_help first, or run the function you’re having issues with using the -? switch. Or simply ask me here.

Have fun 🙂

Upgrading the Vostro 1700 to beyond maximum specs

WARNING: This post is very specific to an old laptop, probably only interesting to anyone looking to upgrade such a machine. I warned you.

I have an unhealthy obsession with “maxing out” machines, especially older ones. It was time for the chunky old Vostro 1700 that now runs Windows 7 instead of the version of Windows XP I installed previously. From various forums I gathered that the laptop’s beefiest configuration from the factory was a Core 2 Duo T9300 CPU and 4 gigs of RAM, available towards the end of its production run.

The Vostro 1700 features an excellent keyboard (for my taste), and a clear and bright full HD matte finish 17″ LCD screen, which is the reason I keep it around. It does Steam in-home streaming reasonably well, and remote desktop to more powerful machines is of course flawless.

Anyway, not going to settle for the factory maximum, I decided to get hold of some memory and a C2D CPU from eBay. The result was a C2D T9500 CPU, which is a tad faster than the T9300, but with similar power requirements, as well as two sticks of Samsung 4GB DDR2 SODIMM ram, which would have been stupid expensive back in the day. These guys estimate well over a hundred dollars, even from eBay, which this machine frankly isn’t worth, but today it’s a mere fraction of that.

Annoyingly, it turns out that even with the most recent BIOS, version A07, 8GB RAM makes the machine hang during initialization. I was, however, able to get 4+2GB to run, after it spent around 10 minutes thinking about it at first boot. For some reason the thing takes ages to boot the first time after a change in the amount of system RAM.

Now for some before and after shots of the Vostro 1700, first with 4 gigs of ram and a T8300 CPU, and then with 6 gigs and the T9500.



Figuring out which session is blocking a query on Microsoft SQL Server (query suspended)

I was recently debugging a case where a customer’s installation was hanging due to a database lock. The issue turned out to be that the database was not set to a default transaction level of READ COMMITTED SNAPSHOT, which the product expects. Regardless, while troubleshooting the issue, I stumbled upon a very useful SQL query that I needed to save somewhere, so here goes.

The following query lists which SQL session is blocked by which other session:

db.name DBName,
OBJECT_NAME(p.OBJECT_ID) BlockedObjectName,
h1.TEXT AS RequestingText,
h2.TEXT AS BlockingTest,
FROM sys.dm_tran_locks AS tl
INNER JOIN sys.databases db ON db.database_id = tl.resource_database_id
INNER JOIN sys.dm_os_waiting_tasks AS wt ON tl.lock_owner_address = wt.resource_address
INNER JOIN sys.partitions AS p ON p.hobt_id = tl.resource_associated_entity_id
INNER JOIN sys.dm_exec_connections ec1 ON ec1.session_id = tl.request_session_id
INNER JOIN sys.dm_exec_connections ec2 ON ec2.session_id = wt.blocking_session_id
CROSS APPLY sys.dm_exec_sql_text(ec1.most_recent_sql_handle) AS h1
CROSS APPLY sys.dm_exec_sql_text(ec2.most_recent_sql_handle) AS h2
ORDER BY tl.request_session_id

…and by following the tree (X blocked by Y, blocked by Z, etc..) one can see which session is responsible for the mess.

Other interesting queries:

sp_who2 Shows sessions connected to the database, how much time they’ve spent, as well as several other stats
select cmd,* from sys.sysprocesses
where blocked > 0
Shows currently blocked queries


Using IPv6 SLAAC with IP forwarding on Debian Stretch (“connect: network is unreachable”, no default gateway)

When configured as a router, for instance for hosting virtual machines with KVM, a Debian machine with a somewhat recent kernel will not listen to router advertisements from others. This makes sense in many cases, but when you have a VM host in your internal network, which only router function is to allow communication to and from the virtual machines it hosts, the default behaviour is less than optimal.

ping6 reports the network as unreachable. Looking at ip -6 route reveals the problem. There’s no “default” line configured.

# ping6 google.com
connect: Network is unreachable
# ip -6 route
2001:4fa0:baaa:1::/64 dev eth0 proto kernel metric 256 expires 86394sec pref medium
fe80::/64 dev eth0 proto kernel metric 256 pref medium

After verifying with tcpdump -i eth0 ip6 that router advertisements were indeed being received, I looked around for a bit and found this post by Andy Smith that explains the problem in detail.

In short, accepting router advertisements is governed by /proc/sys/net/ipv6/conf/$IFACE/accept_ra. These values are documented in ip-sysctl.txt (local archive here).

accept_ra - INTEGER
	Accept Router Advertisements; autoconfigure using them.

	It also determines whether or not to transmit Router
	Solicitations. If and only if the functional setting is to
	accept Router Advertisements, Router Solicitations will be

	Possible values are:
		0 Do not accept Router Advertisements.
		1 Accept Router Advertisements if forwarding is disabled.
		2 Overrule forwarding behaviour. Accept Router Advertisements
		  even if forwarding is enabled.

	Functional default: enabled if local forwarding is disabled.
			    disabled if local forwarding is enabled.

The solution is thus to set accept_ra to 2 to allow for accepting router advertisements even if forwarding is set to 1.

Since I use Shorewall to enable the forwarding, I ended up with a somewhat different config than Andy, but I still use the pre-up statements in /etc/network/interfaces to resolve the issue.

allow-hotplug eth0
iface eth0 inet static
	pre-up echo 2 > /proc/sys/net/ipv6/conf/$IFACE/accept_ra

The rest of the config is only shown for reference. The important part is the pre-up line. Setting this, and then manually writing to /proc/sys/net/ipv6/conf/eth0/accept_ra to apply the configuration immediately, I had a default route within seconds, and ping to Google worked like a charm.

# ip -6 route
2001:4fa0:baaa:1::/64 dev eth0 proto kernel metric 256 expires 86394sec pref medium
fe80::/64 dev eth0 proto kernel metric 256 pref medium
default via fe80::1:1 dev eth0 proto ra metric 1024  expires 50sec hoplimit 64 pref medium
# ping6 -c4 google.com
PING google.com(arn09s11-in-x0e.1e100.net (2a00:1450:400f:807::200e)) 56 data bytes
64 bytes from arn09s11-in-x0e.1e100.net (2a00:1450:400f:807::200e): icmp_seq=1 ttl=55 time=19.0 ms
64 bytes from arn09s11-in-x0e.1e100.net (2a00:1450:400f:807::200e): icmp_seq=2 ttl=55 time=18.6 ms
64 bytes from arn09s11-in-x0e.1e100.net (2a00:1450:400f:807::200e): icmp_seq=3 ttl=55 time=22.3 ms
64 bytes from arn09s11-in-x0e.1e100.net (2a00:1450:400f:807::200e): icmp_seq=4 ttl=55 time=17.8 ms

--- google.com ping statistics ---
4 packets transmitted, 4 received, 0% packet loss, time 3004ms
rtt min/avg/max/mdev = 17.823/19.474/22.359/1.725 ms

Exim4 line length in Debian Stretch – “Mail delivery failed: returning message to sender”

Exim4 introducted new behaviour in the versions included in Debian Stretch. Suddenly, the RFC max line length of 998 characters is enforced, and emails with lines exceeding this length are returned to sender. For the interested, part of the discussion about this feature is over here: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=839147

Anyway, the solution to this issue in Debian is to add the following to /etc/exim4/update-exim4.conf.conf


..then restart Exim4

service exim4 restart

You can test the new configuration by sending an email to ‘root’ with an absurdly long line in it:

$ { echo "Test begins"; for ((i=0; i<1500; i++)); do echo -n "0"; done; echo -e "\nTest ends"; } | mail -s "Very long line" root

If you get that email, you did this correctly 🙂

Understanding mcelog ECC errors – Which stick of RAM is broken?

Before reading this, please note that much of the information in mcelog is hardware dependent. Your mileage may vary.

Memory gone bad

So one of the servers, running an X99-WS/IPMI board from Asus, began putting errors into /var/log/mcelog. Thankfully, they were all the same, telling me the following:

mcelog: failed to prefill DIMM database from DMI data
Hardware event. This is not a software error.
CPU 0 BANK 11 
MISC 90840080008228c ADDR 9ce494000 
TIME 1499161840 Tue Jul 4 09:50:40 2017
MCG status:
MCi status:
Corrected error
MCi_MISC register valid
MCi_ADDR register valid
Transaction: Memory scrubbing error
MemCtrl: Corrected patrol scrub error
STATUS 8c000051000800c2 MCGSTATUS 0
CPUID Vendor Intel Family 6 Model 79

So, what does this mean?

The first few lines tell us this happened from CPU 0 on BANK 11. This wasn’t much help, as the board only has 8 memory banks. It was suggested in #debian on freenode that the high bank number might be due to dual channel memory, but then how do I pinpoint the physical stick?

Enter dmidecode

mcelog tells us a weird bank number, but it has something else that’s vitally important; the address. ADDR 9ce494000 is a memory address on the faulty stick, and dmidecode can tell us which stick is responsible for that address:

# dmidecode -t 20
Handle 0x005E, DMI type 20, 35 bytes
Memory Device Mapped Address
    Starting Address: 0x00800000000
    Ending Address: 0x00BFFFFFFFF
    Range Size: 16 GB
    Physical Device Handle: 0x005D
    Memory Array Mapped Address Handle: 0x0058
    Partition Row Position: 1

This should be the problematic RAM stick, as address 0x009CE494000 is between 0x00800000000 and 0x00BFFFFFFFF. The stick has “Physical Device Handle” 0x005D. dmidecode can show us more information about this handle:

# dmidecode -t 17
Handle 0x005D, DMI type 17, 40 bytes
Memory Device
    Array Handle: 0x0057
    Error Information Handle: Not Provided
    Total Width: 72 bits
    Data Width: 72 bits
    Size: 16384 MB
    Form Factor: RIMM
    Set: None
    Locator: DIMM_B1
    Bank Locator: NODE 1
    Type: DDR4
    Type Detail: Synchronous
    Speed: 2133 MHz
    Manufacturer: Samsung
    Serial Number: 32BFE65D
    Asset Tag: DIMM_B1_AssetTag
    Part Number: M393A2G40DB0-CPB 
    Rank: 2
    Configured Clock Speed: 2133 MHz
    Minimum Voltage: Unknown
    Maximum Voltage: Unknown
    Configured Voltage: Unknown

Here you can look at the Locator, or the Asset Tag fields. Both show the memory slot as DIMM_B1. Now that’s something we can use! Looking in the motherboard manual, available online, one can see where DIMM_B1 is:

So that’s the bad stick, which will be going back to the supplier with an RMA.

Finding duplicate files with Bash

Someone recently asked, in #bash on Freenode, how to find duplicate files with Bash. Several options were suggested, and the user ended up installing and running “fdupes“. However, this sort of thing should be reasonably easy to do using “find” and a few pipes.

As a quick overview, what you want to achieve is to find all the files, list their sizes and names followed by a NUL ($’\0′) separator to allow for wonky filenames. You start with this, as sizes are the quickest values to compare. You then get all the duplicates that have at least one other file with the same size, and discard the rest of them. Then we want to run some sort of a checksum tool. I chose sha1sum. Again we discard the files with no duplicates. And that’s really all there is to it.

Feeling bored one night, I decided to try this. As usual, the whole thing grew entirely out of proportion, and I ended up duplicating most of fdupes’ functionality in Bash. Still, the resulting script is surprisingly fast, looks nice, has a decent help menu and doesn’t rely on anything not found on virtually every GNU/Linux machine out there. It also supports a few things fdupes does not, such as (as of 2017-07-01) null terminated output. Thus it found its way into my toolbox.

Let me present “cdupes”:

$ cdupes
cdupes: no directories specified

  Bash script that's functionally similar to the tool "fdupes"

Usage: cdupes [options] 
    Null termination. Allows piping filenames with weird characters into other tools
    md5sum. Uses md5sum instead of the default sha1sum
    NoHidden. Excludes files which names start with --> . <--
    Checksum. Show checksum of duplicate files
    Omit first match from each group. Useful with -m for cleanup scripts
    Machine readable. No empty lines between groups
    Probably only useful in conjunction with -c or -f
    NoEmpty. Ignore empty files (size 0)
    Permissions. Don't consider files with different owner/group or permission bits as duplicates
    Recurse. For every directory given, follow subdirectories encountered within.
    Size. Show size of duplicate files
    Quiet. Hides progress indicators
    Quiet errors. Errors will not be printed. Does not hide progress indicators

$ cdupes -r /tmp
Files: 21
Same size: 13
Checksum: 13



Duplicate search exited with error status: 1

find: ‘/tmp/nonono’: Permission denied
find: ‘/tmp/root’: Permission denied
find: ‘/tmp/.cathedral’: Permission denied
sha1sum: /tmp/.startup.lock: Permission denied
sha1sum: /tmp/dabba: Permission denied
sha1sum: /tmp/fabba: Permission denied

$ sudo cdupes -rcSq /tmp
2ab06f95377aecc42e5a0e85573a3e7e3efa0961 157286400 /tmp/.fp1
2ab06f95377aecc42e5a0e85573a3e7e3efa0961 157286400 /tmp/fp3
da39a3ee5e6b4b0d3255bfef95601890afd80709 0 /tmp/.startup.lock
da39a3ee5e6b4b0d3255bfef95601890afd80709 0 /tmp/dabba
da39a3ee5e6b4b0d3255bfef95601890afd80709 0 /tmp/fabba
da39a3ee5e6b4b0d3255bfef95601890afd80709 0 /tmp/foo/arse
da39a3ee5e6b4b0d3255bfef95601890afd80709 0 /tmp/test/file-3muc5s
da39a3ee5e6b4b0d3255bfef95601890afd80709 0 /tmp/test/file-8ZjpE9
da39a3ee5e6b4b0d3255bfef95601890afd80709 0 /tmp/root/file-ox8qJb

If you’re interested in trying my script, or including it in your own set of SysOp tools,
you can –> find it here <–
Have fun!