Just had my COVID-19 first vaccination (Pfizer/BioNTech)

Just got back from having my first COVID-19 vaccination. Started queueing at 10:40, pre-screening questions at 10:50, all done by 10:53 then I poked at my phone for 15 minutes while waiting to check I wouldn’t keel over from anaphylactic shock (I didn’t).

I was first notified that I should book an appointment in the form of a text message from sender “GPSurgery” on Monday 22nd February 2021:

Dear MR SMITH,

You have been invited to book your COVID-19 vaccinations.

Please click on the link to book: https://accurx.thirdparty.nhs.uk/…
[Name of My GP Surgery]

The web site presented me with a wide variety of dates and times, the earliest being today, 3 days later, so I chose that. My booking was then confirmed by another text message, and another reminder message was sent yesterday. I assume these text messages were sent by some central service on behalf of my GP whose role was probably just submitting my details.

A very smooth process a 15 minute walk from my home, and I’m hearing the same about the rest of the country too.

Watching social media mentions from others saying they’ve had their vaccination and also looking at the demographics in the queue and waiting room with me, I’ve been struck by how many people have—like me—been called up for their vaccinations quite early unrelated to their age. I was probably in the bottom third age group in the queue and waiting area: I’m 45 and although most seemed older than me, there were plenty of people around my age and younger there.

It just goes to show how many people in the UK are relying on the NHS for the management of chronic health conditions that may not be obviously apparent to those around them. Which is why we must not let this thing that so many of us rely upon be taken away. I suspect that almost everyone reading either is in a position of relying upon the NHS or has nearest and dearest who do.

The NHS gets a lot of criticism for being a bottomless pit of expenditure that is inefficient and slow to embrace change. Yes, healthcare costs a lot of money especially with our ageing population, but per head we spend a lot less than many other countries: half what the US spends per capita or as a proportion of GDP; our care is universal and our life expectancy is slightly longer. In 2017 the Commonwealth Fund rated the NHS #1 in a comparison of 11 countries.

So the narrative that the NHS is poor value for money is not correct. We are getting a good financial deal. We don’t necessarily need to make it perform better, financially, although there will always be room for improvement. The NHS has a funding crisis because the government wants it to have a funding crisis. It is being deliberately starved of funding so that it fails.

The consequences of selling off the NHS will be that many people are excluded from care they need to stay alive or to maintain a tolerable standard of living. As we see with almost every private sector takeover of what were formerly public services, they strip the assets, run below-par services that just about scrape along, and then when there is any kind of downturn or unexpected event they fold and either beg for bailout or just leave the mess in the hands of the government. Either way, taxpayers pay more for less and make a small group of wealthy people even more wealthy.

We are such mugs here in UK that even other countries have realised that they can bid to take over our public services, provide a low standard of service at a low cost to run, charge a lot to the customer and make a hefty profit. Most of our train operating companies are owned by foreign governments.

The NHS as it is only runs as well as it does because the staff are driven to breaking point with an obscene amount of unpaid overtime and workplace stress.

If you’d like to learn some more about the state of the NHS in the form of an engaging read then I recommend Adam Kay’s book This is Going to Hurt: Secret Diaries of a Junior Doctor. It will make you laugh, it will make you cry and if you’ve a soul it will make you angry. Also it may indelibly sear the phrase “penis degloving injury” into your mind.

Do not accept the premise that the NHS is too expensive.

If the NHS does a poor job (and it sometimes does), understand that underfunding plays a big part.

Privatising any of it will not improve matters in any way, except for a very small number of already wealthy people.

Please think about this when you vote.

Intel may need me to sign an NDA before I can know the capacity of one of their SSDs

Apologies for the slightly clickbaity title! I could not resist. While an Intel employee did tell me this, they are obviously wrong.

Still, I found out some interesting things that I was previously unaware of.

I was thinking of purchasing some “3.84TB” Intel D3-S4610 SSDs for work. I already have some “3.84TB” Samsung SM883s so it would be good if the actual byte capacity of the Intel SSDs were at least as much as the Samsung ones, so that they could be used to replace a failed Samsung SSD.

To those with little tech experience you would think that two things which are described as X TB in capacity would be:

  1. Actually X TB in size, where 1TB = 1,000 x 1,000 x 1,000 x 1,000 bytes, using powers of ten SI prefixes. Or;
  2. Actually X TiB in size, where 1TiB = 1,024 x 1,024 x 1,024 x 1,024 bytes, using binary prefixes.

…and there was a period of time where this was mostly correct, in that manufacturers would prefer something like the former case, as it results in larger headline numbers.

The thing is, years ago, manufacturers used to pick a capacity that was at least what was advertised (in powers of 10 figures) but it wasn’t standardised.

If you used those drives in a RAID array then it was possible that a replacement—even from the same manufacturer—could be very slightly smaller. That would give you a bad day as you generally need devices that are all the same size. Larger is okay (you’ll waste some), but smaller won’t work.

So for those of us who like me are old, this is something we’re accustomed to checking, and I still thought it was the case. I wanted to find out the exact byte capacity of this Intel SSD. So I tried to ask Intel, in a live support chat.

Edgar (22/02/2021, 13:50:59): Hello. My name is Edgar and I’ll be helping you today.

Me (22/02/2021, 13:51:36): Hi Edgar, I have a simple request. Please could you tell me the exact byte capacity of a SSD-SSDSC2KG038T801 that is a 3.84TB Intel D3-S4610 SSD

Me (22/02/2021, 13:51:47): I need this information for matching capacities in a RAID set

Edgar (22/02/2021, 13:52:07): Hello, thank you for contacting Intel Technical Support. It is going to be my pleasure to help you.

Edgar (22/02/2021, 13:53:05): Allow me a moment to create a ticket for you.

Edgar (22/02/2021, 13:57:26): We have a calculation to get the decimal drive sectors of an SSD because the information you are asking for most probably is going to need a Non-Disclousre Agreement (NDA)

Yeah, an Intel employee told me that I might need to sign an NDA to know the usable capacity of an SSD. This is obviously nonsense. I don’t know whether they misunderstood and thought I was asking about the raw capacity of the flash chips or what.

Me (22/02/2021, 13:58:15): That seems a bit strange. If I buy this drive I can just plug it in and see the capacity in bytes. But if it’s too small then that is a wasted purchase which would be RMA’d

Edgar (22/02/2021, 14:02:48): It is 7,500,000,000

Edgar (22/02/2021, 14:03:17): Because you take the size of the SSD that is 3.84 in TB, in Byte is 3840000000000

Edgar (22/02/2021, 14:03:47): So we divide 3840000000000 / 512 which is the sector size for a total of 7,500,000,000 Bytes

Me (22/02/2021, 14:05:50): you must mean 7,500,000,000 sectors of 512byte, right?

Edgar (22/02/2021, 14:07:45): That is the total sector size, 512 byte

Edgar (22/02/2021, 14:08:12): So the total sector size of the SSD is 7,500,000,000

Me (22/02/2021, 14:08:26): 7,500,000,000 sectors is only 3,750GB so this seems rather unlikely

The reason why this seemed unlikely to me is that I have never seen an Intel or Samsung SSD that was advertised as X.Y TB capacity that did not have a usable capacity of at least X,Y00,000,000,000 bytes. So I would expect a “3.84TB” device to have at least 3,840,000,000,000 bytes of usable capacity.

Edgar was unable to help me further so the support chat was ended. I decided to ask around online to see if anyone actually had one of these devices running and could tell me the capacity.

Peter Corlett responded to me with:

As per IDEMA LBA1-03, the capacity is 1,000,194,048 bytes per marketing gigabyte plus 10,838,016 bytes. A marketing terabyte is 1000 marketing gigabytes.

3840 * 1000194048 + 10838016 = 3840755982336. Presumably your Samsung disk has that capacity, as should that Intel one you’re eyeing up.

My Samsung ones do! And every other SSD I’ve checked obeys this formula, which explains why things have seemed a lot more standard recently. I think this might have been standardised some time around 2014 / 2015. I can’t tell right now because the IDEMA web site is down!

So the interesting and previously unknown to me thing is that storage device sizes are indeed standardised now, albeit not to any sane definition of the units that they use.

What a relief.

Also that Intel live support sadly can’t be relied upon to know basic facts about Intel products. 🙁

Booting the CentOS/RHEL installer under Xen PVH mode

CentOS/RHEL and Xen ^

As of the release of CentOS 8 / RHEL8, Red Hat disabled kernel support for running as a Xen PV or PVH guest, even though such support is enabled by default in the upstream Linux kernel.

As a result—unlike with all previous versions of CentOS/RHEL—you cannot boot the installer in Xen PV or PVH mode. You can still boot it in Xen HVM mode, or under KVM, but that is not very helpful if you don’t want to run HVM or KVM.

At BitFolk ever since the release of CentOS 8 we’ve had to tell customers to use the Rescue VM (a kind of live system) to unpack CentOS into a chroot.

Fortunately there is now a better way.

Credit ^

This method was worked out by Jon Fautley. Jon emailed me instructions and I was able to replicate them. Several people have since asked me how it was done and Jon was happy for me to write it up, but this was all worked out by Jon, not me.

Overview ^

The basic idea here is to:

  1. take the installer initrd.img
  2. unpack it
  3. shove the modules from a Debian kernel into it
  4. repack it
  5. use a Debian kernel and this new frankeninitrd as the installer kernel and initrd
  6. switch the installed OS to kernel-ml package from ELRepo so it has a working kernel when it boots

Detailed process ^

I’ll go into enough detail that you should be able to exactly replicate what I did to end up with something that works. This is quite a lot but it only needs to be done each time the real installer initrd.img changes, which isn’t that often. The resulting kernel and initrd.img can be used to install many guests.

Throughout the rest of this article I’ll refer to CentOS, but Jon initially made this work for RHEL 8. I’ve replicated it for CentOS 8 and will soon do so for RHEL 8 as well.

Extract the CentOS initrd.img ^

You will find this in the install ISO or on mirrors as images/pxeboot/initrd.img.

$ mkdir /var/tmp/frankeninitrd/initrd
$ cd /var/tmp/frankeninitrd/initrd
$ xz -dc /path/to/initrd.img > ../initrd.cpio
$ # root needed because this will do some mknod/mkdev.
$ sudo cpio -idv < ../initrd.cpio

Copy modules from a working Xen guest ^

I’m going to use the Xen guest that I’m doing this on, which at the time of writing is a Debian buster system running kernel 4.19.0-13. Even a system that is not currently running as a Xen guest will probably work, as they usually have modules available for everything.

At the time of writing the kernel version in the installer is 4.18.0-240.

If you’ve got different, adjust filenames accordingly.

$ sudo cp -r /lib/modules/4.19.0-13-amd64 lib/modules/
$ # You're not going to use the original modules
$ # so may as well delete them to save space.
$ sudo rm -vr lib/modules/4.18*

Add dracut hook to copy fs modules ^

$ cat > usr/lib/dracut/hooks/pre-pivot/99-move-modules.sh <<__EOF__
#!/bin/sh
 
mkdir -p /sysroot/lib/modules/$(uname -r)/kernel/fs
rm -r /sysroot/lib/modules/4.18*
cp -r /lib/modules/$(uname -r)/kernel/fs/* /sysroot/lib/modules/$(uname -r)/kernel/fs
cp /lib/modules/$(uname -r)/modules.builtin /sysroot/lib/modules/$(uname -r)/
depmod -a -b /sysroot
 
exit 0
__EOF__
$ chmod +x usr/lib/dracut/hooks/pre-pivot/99-move-modules.sh

Repack initrd ^

This will take a really long time because xz -9 is sloooooow.

$ sudo find . 2>/dev/null | \
  sudo cpio -o -H newc -R root:root | \
  xz -9 --format=lzma > ../centos8-initrd.img

Use the Debian kernel ^

Put the matching kernel next to your initrd.

$ cp /boot/vmlinuz-4.19.0-13-amd64 ../centos8-vmlinuz
$ ls -lah ../centos*
-rw-r--r-- 1 andy andy  81M Feb  1 04:43 ../centos8-initrd.img
-rw-r--r-- 1 andy andy 5.1M Feb  1 04:04 ../centos8-vmlinuz

Boot this kernel/initrd as a Xen guest ^

Copy the kernel and initrd to somewhere on your dom0 and create a guest config file that looks a bit like this:

name       = "centostest"
# CentOS 8 installer requires at least 2.5G RAM.
# OS will run with a lot less though.
memory     = 2560
vif        = [ "mac=00:16:5e:00:02:39, ip=192.168.82.225, vifname=v-centostest" ]
type       = "pvh"
kernel     = "/var/tmp/frankeninitrd/centos8-vmlinuz"
ramdisk    = "/var/tmp/frankeninitrd/centos8-initrd.img"
extra      = "console=hvc0 ip=192.168.82.225::192.168.82.1:255.255.255.0:centostest:eth0:none nameserver=8.8.8.8 inst.stage2=http://www.mirrorservice.org/sites/mirror.centos.org/8/BaseOS/x86_64/os/ inst.ks=http://example.com/yourkickstart.ks"
disk       = [ "phy:/dev/vg/centostest_xvda,xvda,w",
               "phy:/dev/vg/centostest_xvdb,xvdb,w" ]

Assumptions in the above:

  • vif and disk settings will be however you usually do that.
  • “extra” is for the kernel command line and here gives the installer static networking with the ip=IP address::default gateway:netmask:hostname:interface name:auto configuration type option.
  • inst.stage2 here goes to a public mirror but could be an unpacked installer iso file instead.
  • inst.ks points to a minimal kickstart file you’ll have to create (see below).

Minimal kickstart file ^

This kickstart file will:

  • Automatically wipe disks and partition. I use xvda for the OS and xvdb for swap. Adjust accordingly.
  • Install only minimal package set.
  • Switch the installed system over to kernel-ml from EPEL.
  • Force an SELinux autorelabel at first boot.

The only thing it doesn’t do is create any users. The installer will wait for you to do that. If you want an entirely automated install just add the user creation stuff to your kickstart file.

url --url="http://www.mirrorservice.org/sites/mirror.centos.org/8/BaseOS/x86_64/os"
text
 
# Clear all the disks.
clearpart --all --initlabel
zerombr
 
# A root filesystem that takes up all of xvda.
part /    --ondisk=xvda --fstype=xfs --size=1 --grow
 
# A swap partition that takes up all of xvdb.
part swap --ondisk=xvdb --size=1 --grow
 
bootloader --location=mbr --driveorder=xvda --append="console=hvc0"
firstboot --disabled
timezone --utc Etc/UTC --ntpservers="0.pool.ntp.org,1.pool.ntp.org,2.pool.ntp.org,3.pool.ntp.org"
keyboard --vckeymap=gb --xlayouts='gb'
lang en_GB.UTF-8
skipx
firewall --enabled --ssh
halt
 
%packages
@^Minimal install
%end 
 
%post --interpreter=/usr/bin/bash --log=/root/ks-post.log --erroronfail
 
# Switch to kernel-ml from EPEL. Necessary for Xen PV/PVH boot support.
rpm --import https://www.elrepo.org/RPM-GPG-KEY-elrepo.org
yum -y install https://www.elrepo.org/elrepo-release-8.el8.elrepo.noarch.rpm
yum --enablerepo=elrepo-kernel -y install kernel-ml
yum -y remove kernel-tools kernel-core kernel-modules
 
sed -i -e 's/DEFAULTKERNEL=.*/DEFAULTKERNEL=kernel-ml/' /etc/sysconfig/kernel
grub2-mkconfig -o /boot/grub2/grub.cfg
 
# Force SELinux autorelabel on first boot.
touch /.autorelabel
%end

Launch the guest ^

$ sudo xl create -c /etc/xen/centostest.conf

Obviously this guest config can only boot the installer. Once it’s actually installed and halts you’ll want to make a guest config suitable for normal booting. The kernel-ml does work in PVH mode so at BitFolk we use pvhgrub to boot these.

A better way? ^

The actual modifications needed to the stock installer kernel are quite small: just enable CONFIG_XEN_PVH kernel option and build. I don’t know the process to build a CentOS or RHEL installer kernel though, so that wasn’t an option for me.

If you do know how to do it please do send me any information you have.

If you’re running Ubuntu and/or using snaps, look into CVE-2020-27348

I was reading an article about CVE-2020-27348 earlier, which is quite a nasty bug affecting a lot of snap packages.

My desktop runs Ubuntu 18.04 at the moment, and so does my partner’s laptop. I also have a Debian buster laptop but I’ve never installed snapd there. So it’s just my desktop and my partner’s laptop I’m concerned about.

If you run Ubuntu 20.04 or later I think there’s probably more concern, as I understand the software centre offers snap versions of things by default.

Anyway, I couldn’t recall ever installing a snap on purpose on my desktop except for a short while ago when I intentionally installed signal-desktop. But in fact I have quite a few snaps installed.

$ snap list
Name                  Version                     Rev    Tracking         Publisher     Notes
core                  16-2.48.2                   1058   latest/stable    canonical✓    core
core18                20201210                    1944   latest/stable    canonical✓    base 
gnome-3-26-1604       3.26.0.20200529             100    latest/stable/…  canonical✓    -
gnome-3-28-1804       3.28.0-19-g98f9e67.98f9e67  145    latest/stable    canonical✓    -
gnome-3-34-1804       0+git.3556cb3               66     latest/stable    canonical✓    -
gnome-calculator      3.38.0+git7.c840c69c        826    latest/stable/…  canonical✓    -
gnome-characters      v3.34.0+git9.eeab5f2        570    latest/stable/…  canonical✓    -
gnome-logs            3.34.0                      100    latest/stable/…  canonical✓    -
gnome-system-monitor  3.36.0-12-g35f88a56d7       148    latest/stable/…  canonical✓    -
gtk-common-themes     0.1-50-gf7627e4             1514   latest/stable/…  canonical✓    -
signal-desktop        1.39.5                      345    latest/stable    snapcrafters  -

I don’t know why gnome-calculator is there. It doesn’t appear to be the binary that’s run when I start the calculator.

So are any of them a security risk? Well…

$ grep -l \$LD_LIBRARY_PATH /snap/*/current/snap/snapcraft.yaml
/snap/gnome-calculator/current/snap/snapcraft.yaml
/snap/gnome-characters/current/snap/snapcraft.yaml
/snap/gnome-logs/current/snap/snapcraft.yaml
/snap/gnome-system-monitor/current/snap/snapcraft.yaml

Those are all the snaps on my system which include the value of the (empty) environment variable LD_LIBRARY_PATH, so are likely vulnerable to this.

But does this really end up with an empty item in the LD_LIBRARY_PATH list?

$ which gnome-system-monitor 
/snap/bin/gnome-system-monitor
$ gnome-system-monitor &
$ pgrep -f gnome-system-monitor
8259
$ tr '\0' '\n' < /proc/8259/environ | grep ^LD_LIBR | grep -q :: && echo "oh dear"
oh dear

Yes it really does.

(The tr is necessary above because the /proc/*/environ file is a NUL-separated string, so that modifies it to be one variable per line, then looks for the LD_LIBRARY_PATH line, and checks if it has an empty entry ::)

So yeah, my gnome-system-monitor is a local code execution vector.

As are my gnome-characters, gnome-logs and that gnome-calculator if I ever uninstall the non-snap version.

That CVE seems to have been published on 3 December 2020. I hope that the affected snaps will be fixed soon.

I don’t like that the CVE says the impact is:

If a user were tricked into installing a malicious snap or downloading a malicious library, under certain circumstances an attacker could exploit this to affect strict mode snaps that have access to the library and were launched from the directory containing the library.

My first thought upon reading is, “I’m safe, I haven’t been tricked into downloading any malicious snaps!” But I do have snaps that aren’t malicious, they are just insecure. The hardest part of the exploit is indeed getting a malicious file (a library) into my filesystem in a directory where I will run a snap from.

Fun With SpamAssassin Meta Rules

I’ve got a ticketing system. Let’s say you open a ticket by emailing support@example.com. You then get an automated response confirming that you’ve opened a ticket, and on my side people get bothered by a notification about this support ticket that needs attention.

A problem here is that absolutely anyone or anything emailing that will open a ticket. And it’s pretty easy to find that email address.

As a result lots of scum of the earthenterprising individuals seem to be passing that email address around to other enterprising individuals who decide to add it to their email marketing mailshots.

A reasonable response to this would perhaps be to move away from email to a web form, and put it behind a login so that only existing, authenticated customers could submit new tickets. Thing is, I still have to have a way for previously-unknown people to create tickets by email, and I kind of like email. So I persevere.

One thing I can do though is block all kinds of newsletters. There is no scenario where people who send newsletters should be trying to open support tickets. I’m prepared to disallow any email from MailJet or SendGrid being sent to my ticketing system, for example.

But how to do it?

Well, I am already able to identify emails from MailJet and SendGrid because I use the ASN plugin. This inserts a header in the email to say which Autonomous System it came from.

MailJet’s ASN is 200069 and SendGrid’s is 11377. I know that because I’ve seen mail from them before, and the ASN plugin put a header in with those numbers.

You can add some custom rules to match mails from these ASNs:

header   LOCAL_ASN_MAILJET X-ASN =~ /\b200069\b/
score    LOCAL_ASN_MAILJET 0.001
describe LOCAL_ASN_MAILJET Sent by MailJet (ASN200069)

What this will do is check the header that the ASN plugin added and if it matches it will add this label LOCAL_ASN_MAILJET with a score of 0.001 to the list of SpamAssassin scores.

Scores that are very close to zero (but not actually zero!) are typically used just to annotate an email. You can’t use zero exactly because that disables the rule entirely.

Now, if you really didn’t want any email from MailJet at all you could crank that score up and it would all be rejected. But my users do actually get quite a lot of wanted email from the likes of MailJet and SendGrid. These senders are sadly too big to block. They know this, and this probably contributes to their noted preference for taking spammers’ money, but that is a rant for another day.

Back to the original goal: I only want to reject mail from these companies if it is destined for my ticketing system. So how to identify mail that’s for the support queue? Well that’s pretty simple:

header   LOCAL_TO_SUPPORT ToCc:addr =~ /^support\@example\.com$/i
score    LOCAL_TO_SUPPORT 0.001
describe LOCAL_TO_SUPPORT Recipient is support queue

This checks just the address part(s) of the To and Cc headers to see if any match support@example.com. The periods (‘.’) and the at symbol (‘@’) need escaping because this is a Perl regular expression. If there’s a match then the LOCAL_TO_SUPPORT tag will be added.

Now all that remains is to make a new rule that only fires if both of these conditions are true, and assigns a real score to that:

meta     LOCAL_MAILSHOT_TO_SUPPORT (LOCAL_TO_SUPPORT && (LOCAL_ASN_MAILJET || LOCAL_ASN_SENDGRID))
score    LOCAL_MAILSHOT_TO_SUPPORT 10.0
describe LOCAL_MAILSHOT_TO_SUPPORT Mailshot sent to support queue

There. Now the support queue will never get emails from these companies, but the rest of my users still can.

Of course you don’t have to match those mails by ASN. There are many other indicators of senders that just shouldn’t be opening support tickets, and if you can find any other sort of rule that matches them reliably then you can chain that with other rules that identify the support queue recipient.

Another way to do it would be to run the support queue as its own SpamAssassin user with its own per-user rules. I have a fairly simple SpamAssassin setup though with only a global set of rules so I didn’t want to do that just for this.

Starting services only when the network is ready on Debian/systemd

TL;DR: ^

  • Make sure that whatever configures your network supports network-online.target
  • Override the service unit to have Wants=network.target network-online.target and the same for After=

Overview ^

Sometimes you only want services to start up once there is a network configured. Most network services can handle the situation where there is initially no network, waiting until the network appears, because this is a very common situation.

Other services though may not in themselves be expecting to use the network, and so have never thought about it. Also a great thing about open source software is that it tends to be very composable, so it’s not possible to predict the ways that people will use combinations of software.

The problem ^

systemd will tend to start things as soon as it can. If your service is not configured to wait for the network that means it will most likely be started up before the network exists. If your service then tries to do something that requires a network it will receive an error, which it may not be prepared to handle.

A concrete example: ulogd2 ^

A real life example for me is ulogd2. ulogd2 allows your firewall rules to log things in a variety of ways, in incredible detail.

Most of the ways people configure it involve just logging to the local filesystem, so it doesn’t actually require the network to be configured first.

The default systemd configuration in Debian buster for the ulogd2 service looks like this:

$ sudo systemctl cat ulogd2.service
[Unit]
Description=Netfilter Userspace Logging Daemon
Documentation=man:ulogd(8)
 
[Service]
Type=forking
PIDFile=/run/ulog/ulogd.pid
ExecStart=/usr/sbin/ulogd --daemon --uid ulog --pidfile /run/ulog/ulogd.pid
 
[Install]
Alias=ulogd.service
WantedBy=multi-user.target

As you can see there’s nothing in there that says to wait for a network.

I use a database plugin for ulogd2 that makes it log to a (remote) database. As a consequence as soon as it starts up it tries to establish a database connection, immediately fails as there is no route to any remote host, retries a few times and then bails out.

Most of the time it exhausts its retries before the network is up, so the result is that the service is in a failed state. Simply manually starting the service (or having config management do it) resolves that, but that’s a mess.

Ideally I don’t want systemd to start ulogd2 until there is a network.

Wants=network-online.target” mate. Job done. No! ^

If like me you know just enough about systemd to be dangerous, you figure that what you want to do is add something like this to the [Unit] section of the service unit file:

[Unit]
…
Wants=network.target network-online.target
After=network.target network-online.target

This is only part of the correct solution. If you do only this then you’ll probably find that nothing actually changes.

About network-online.target ^

The thing about the network-online target is that it doesn’t exist unless you’re using a “modern” method of bringing up your networking, like NetworkManager or systemd-networkd.

If you’re not doing that then systemd works out that the network-online target can never be reached and ignores it as a Want.

I’m using ifupdown on servers as it still does everything I need it to. To make ifupdown support the network-online target on Debian, you should enable the ifupdown-wait-online service:

$ sudo systemctl enable ifupdown-wait-online.service

This will inject the network-online “target reached” state when every interface that is marked as “auto” in /etc/network/interfaces is up.

Editing a service file ^

The temptation now may be to edit the ulogd2 service file that’s under /lib/systemd/system/ to contain the Want/After bits.

That will work but it isn’t the correct way because if there is a package update then your changes will be overwritten.

A better way is to place a new service file into /etc/systemd/system/. That will entirely override the distributed copy. The obvious downside there is that if there’s an improvement to the packaged service file then you’ll never use it, as you’ve entirely overridden it with your own file.

Overrides to the rescue ^

The best way is to use an override file, and the easiest way to do that is with systemctl edit:

$ sudo systemctl edit ulogd2
[your favourite editor starts]
[Unit]
Wants=network.target network-online.target
After=network.target network-online.target

Check your changes took effect:

$ sudo systemctl cat ulogd2.service
# /lib/systemd/system/ulogd2.service
[Unit]
Description=Netfilter Userspace Logging Daemon
Documentation=man:ulogd(8)
 
[Service]
Type=forking
PIDFile=/run/ulog/ulogd.pid
ExecStart=/usr/sbin/ulogd --daemon --uid ulog --pidfile /run/ulog/ulogd.pid
 
[Install]
Alias=ulogd.service
WantedBy=multi-user.target
 
# /etc/systemd/system/ulogd2.service.d/override.conf
[Unit]
Wants=network.target network-online.target
After=network.target network-online.target

Note that this shows you where the files actually are. That makes it easy to distribute this through config management.

Debian-installer, mdadm configuration and the Bad Blocks Controversy

Updates! ^

Since this was posted on 2020-09-13 there was some interest in the comments and on Hacker News and I learned some things which required updates. I’ve tried to indicate them with struck out text.

Of particular note is the re-add method of removing BBLs.

MD and mdadm ^

MD is the Linux kernel driver that is used for running software RAID arrays. mdadm is the software that you run to manage MD devices. They are both part of the same project.

First, about the Bad Blocks List ^

Since about 2010, MD has had a bad blocks log (BBL) feature. When it fails to read from an underlying device it will (sometimes?) mark that block as bad and read the correct data from a different device, and then forever more redirect reads away from those bad blocks. This feature defaults to being on.

One problem with this feature is that read errors can occur for many reasons besides permanent failure of part of a storage device. For example, it could be a failure of the backplane or controller that causes many read errors on multiple devices, or the devices could be reached over a network of some sort and temporary network problems could propagate errors.

Even if the particular part of the device is unreadable, the operating system is supposed to try to write the correct data over the top. This write will either clear the problem or else be redirected to a spare sector on the drive by the drive’s firmware. The operating system is not supposed to be taking on this role, the drives are, and when the drives fail to do so then the redundancy of the array is supposed to save the day.

Even worse, there are apparently bugs somewhere in the BBL code that cause a device’s BBL to be copied onto a new device when the array is rebuilt or a device replaced. Clearly it does not make sense for a new device to get a copy of another device’s BBL because they are inherently a per-device thing. So far there has been no successful intentional reproduction of this, only people unwittingly hitting it at the worst possible moments. It has been reproduced that adding or replacing a device results in a BBL being copied. I am not aware of a formal bug report for this yet.

mdadm doesn’t even try particularly hard to warn you if a new bad block is found. Unlike when a device fails, it doesn’t send you an email. The MD driver writes in the syslog about the bad block(s). There’s also no change to /proc/mdstat. You have to examine some files in sysfs.

As a result the current situation is that:

No one seems to have made any progress on fixing any of this in 10 years.

Doing something about it ^

I’ll say right now that this story doesn’t (yet?) have a satisfying ending.

I’ve been aware of the “Bad Blocks Controversy” for about 5 years but I haven’t ever personally experienced any problems and it was always at the bottom of my list to look at. Roy’s recent thread spurred me into deciding that in future no MD array I created would have a BBL.

I also took the opportunity to deploy Sarah Newman’s Ansible role which checks all array components have an empty BBL. None of BitFolk‘s array components currently have any entries in their BBLs – phew!

Removing an existing BBL ^

Currently the only way to remove a BBL from an array component is to stop the array and then assemble it with an argument like this:

There are two ways to remove the BBL from the devices of existing arrays.

Fail and re-add each device with update

It doesn’t seem to be documented anywhere, but you can fail a device out of an array and re-add it with an update to remove the BBL on that device, like this:

# mdadm --fail /dev/md0 /dev/sdb1 \
        --remove /dev/sdb1 \
        --re-add /dev/sdb1 \
        --update=no-bbl
mdadm: set /dev/sdb1 faulty in /dev/md0                                              
mdadm: hot removed /dev/sdb1 from /dev/md0                   
mdadm: re-added /dev/sdb1

This will only work if your array has a bitmap, otherwise it will refuse to re-add. Most arrays do get a bitmap, but small arrays won’t by default. Fortunately you can easily add a bitmap like this:

# mdadm --grow --bitmap=internal /dev/md0

The downside of this approach is that your array will have reduced redundancy while it rebuilds. It should rebuild pretty quickly though as the bitmap will cause only changed parts to be rewritten.

(This won’t work if a BBL currently has any entries)

Stop the array and assemble again with update

The other way to remove BBL from devices is to stop the array and assemble it manually like this:

# mdadm --assemble /dev/mdX --update=no-bbl

The big problem with this is that stopping the array obviously causes downtime for whatever is using it. If your root filesystem is on an MD array (and why wouldn’t it be, if you use MD?) then that means the entire server, and you’re having to do this from sort of rescue environment.

I have suggested that a config option be added to remove a BBL on assembly, so that this will happen the next time the machine is rebooted. This does not appear to have provoked any interest.

This method is quicker since it operates on all devices and doesn’t require a rebuild, but personally I usually find downtime more painful so I’d be inclined to schedule an “at-risk” maintenance window and do it the re-add way.

Avoiding the BBL at creation time ^

So if the BBL cannot be easily removed, at least it can be prevented from ever existing, right? When Neil Brown, the previous MD maintainer, was asked in 2016 if the feature could be defaulted to off, Neil said that putting this in the config file was as good as that:

CREATE bbl=no

The thing is, it’s not as good as disabling it by default when you consider what many users’ experience is of running the mdadm command: they don’t run mdadm, something else runs it for them. I’d go as far as to say that the majority of uses of mdadm are done by helper scripts and installers, not by human beings.

If it’s a program that is running mdadm for you then you are going to have to find out how to set that mdadm.conf before it reads it.

Take for example my own process of installing Debian. I do it by booting the Debian Installer by PXE. I have some pre-seeding done to answer a lot of the installer questions, but actually I do still do the disk partitioning stage in the installer’s text interface.

So there I was thinking this is actually going to be quite simple, because the Debian Installer is really lovely about letting you execute a shell and poke around. Surely all I am going to need to do is open a shell once and edit /etc/mdadm/mdadm.conf and then go back into the mdcfg menu and carry on, right? Oh dear me no.

You can read the details of my wild ride that involved me uploading a binary of strace into the d-i to run mdadm under to work out what was going on, but just the relevant discoveries are in this article for those who’d rather not.

mdadm in d-i uses a config file at /tmp/mdadm.conf

After quite a bit of confusion over why even arrays I created manually with the mdadm command in the d-i shell still had a BBL, I discovered that the mdadm binary in d-i is compiled to have its config at /tmp/mdadm.conf. I don’t know why, but probably there is a good reason.

(At this point a number of people responded, “that’s because everything else will be set read-only.” That’s not the case with debian-installer which runs entirely off of a tmpfs. It’s all writeable.)

So just make the edit to /tmp/mdadm.conf then?

Oh ho ho no. Every time you go into the MD configuration section (mdcfg) it clobbers its own /tmp/mdadm.conf, and you can’t get to the “execute a shell” option without returning to the MD configuration section.

If you’re on something with multiple virtual consoles (like if you’re sitting in front of a conventional PC) then you could switch to one of those after you’ve entered the MD configuration part and modify /tmp/mdadm.conf then. I don’t have that option because I’m on a serial console.

I thought I didn’t have that option because I’m on a serial console, but it was pointed out to me that when the Debian installer detects it’s running in a serial console it runs itself under GNU Screen. So, by using the usual screen commands of ctrl+a n or ctrl+a p, one can switch backwards and forwards through the different virtual consoles. Neat!

There is also an earlier option to load an installer component that enables one to continue the installation process over SSH. If you select that then you can SSH in to the running installer system so if you do that after you’ve entered the MD configuration bit in your main console then I guess you can then edit the config file and continue.

By one of those methods of getting a shell, after you’ve already entered the array configuration part but before you’ve actually created any arrays, I think you could edit /tmp/mdadm.conf to have “CREATE bbl=no” and the installer’s mdadm binary would respect that when you switch back.

Alternatively you could just use the shell to create your arrays instead of using the Ddebian installer to do it. If it’s a simple case where you’ve just got an sda and an sdb disk identically partitioned and you want to make a bunch of arrays on them, it can be a fairly legible shell session like:

~ # mkdir -vp /etc/mdadm && echo "CREATE bbl=no" > /etc/mdadm/mdadm.conf
~ # for part in 1 2 3 5; do \
      mdadm --create \
            -v \
            --config=/etc/mdadm/mdadm.conf \
            /dev/md${part} \
            --level=1 \
            --raid-devices=2 \
            /dev/sd[ab]${part}; \
    done

Do not try this until you understand exactly what it is doing.

It iterates the list 1, 2, 3, 5 (I use the 4th partition for something else) and makes arrays called mdX out of sdaX and sdbX. The mdadm binary is forced to use our config file that disables creation of a BBL.

You can verify that a BBL does not exist on any of the array components like this:

~ # mdadm --examine-badblocks /dev/sda1
No bad-blocks list configured on /dev/sda1

You should get identical output for every component. If a component did have a BBL it would output something like this:

~ # mdadm --examine-badblocks /dev/sda1
Bad-blocks list is empty in /dev/sda1

You can then exit the d-i shell and go back to the disk partitioning section. You won’t need the MD configuration part now but even if you do go into it, it should detect all your manually-created arrays.

How to make progress? ^

All of this isn’t great but at least it’s fairly easy to pause the Debian installer and take some manual action. I suspect users of other Linux distributions may not be so lucky, and so I too think it would be a good idea if this buggy feature was disabled by default, or at least if there were a way to tell mdadm to remove the BBL on assembly.

In fact I would very much like to be able to tell it to remove the BBL on assembly so that I can disable the BBL feature on all my existing servers.

mdadm actually gets called by udev from inside the initramfs in incremental assembly mode, so I think the incremental assembly code needs to look in the config file for this “remove all the BBLs” directive and do it then during assembly as if update=no-bbl had been specified on a command line.

It should be possible to write a script that:

  1. Looks in /sys/block/md* to find device components of all arrays.
  2. Checks each one to see if it has a BBL.
  3. If any are found, add a bitmap if necessary.
  4. Do the fail/remove/re-add trick on each one in turn, waiting for the array to go back into sync each time.

i.e. it should be possible to automate this and run it at the end of an install so the entire install process can remain automated, or run it on a host any time after it’s been provisioned.

Recovering From an Exif Disaster

The Discovery ^

Sometime in late December (2019) I noticed that when I clicked on a tag in Shotwell, the photo management software that I use, it was showing either zero or hardly any matching photos when I knew for sure that there should be many more.

(When I say “tag” in this article it’s mostly going to refer to the type of tags you generally put on an image, i.e. the tags that identify who or what is in the image, what event it is associated with, the place it was taken etc. Images can have many different kinds of tags containing all manner of metadata, but for avoidance of doubt please assume that I don’t mean any of those.)

I have Shotwell set to store the tags in the image files themselves, in the metadata. There is a standard for this called Exif. What seems to have happened is that Shotwell had removed a huge number of tags from the files themselves. At the time of discovery I had around 15,500 photos in my collection and it looked like the only way to tell what was in them would be by looking at them. Disaster.

Here follows some notes about what I found out when trying to recover from this situation, in case it si ever useful for anyone.

Shotwell still had a visible tag hierarchy, so I could for example click on the “Pets/Remy” tag, but this brought up only one photo that I took on 14 December 2019. I’ve been taking photos of Remy for years so I knew there should be many more. Here’s Remy.

Remy at The Avenue Ealing Christmas Fair, December 2019
Remy at The Avenue Ealing Christmas Fair

Luckily, I have backups.

Comparing Good and Bad Copies of a Photo ^

I knew this must have happened fairly recently because I’d have noticed quite quickly that photos were “missing”. I had a look for a recent photo that I knew I’d tagged with a particular thing, and then looked in the backups to see when it was last modified.

As an example I found a photo that was taken on 30 October 2019 that should have been tagged “Pets/Violet” but no longer was. It had been modified (but not by me) on 7 December 2019.

A broken photo of Violet
A broken photo of Violet

(Sorry about the text-as-images; I’m reconstructing this series of events from a Twitter thread, where things necessarily had to be posted as screenshots.)

What the above shows is that the version of the photo that existed on 30 October 2019 had the tags “Pets“, “Edna“, and “Violet” but then the version that was written on 7 December 2019 lost the “Violet” tag.

Here I used the exiftool utility to display EXIF tags from the photo files. You can do that like this:

$ exiftool -s $filename

Using egrep I limited this to the tag keys “Subject“, “Keywords“, and “TagsListLastKeywordXMP” but this was a slight mistake: “TagsListLastKeywordXMP” was actually a typo, is totally irrelevant and should be ignored.

Subject” and “Keywords” were always identical for any photo I examined and contained the flattened list of tags. For example, in Shotwell that photo originally had the tags:

  • Pets/Edna
  • Pets/Violet

It seems that Shotwell flattens that to:

  • Pets
  • Edna
  • Violet

and then stores it in “Subject” and “Keywords“.

The tags with hierarchy are actually in the key “TagsList” like:

  • Pets
  • Pets/Edna
  • Pets/Violet

Fixing One Photo ^

I tested stuffing the tag “Violet” back in to this file under the keys “Subject” and “Keywords“:

$ exiftool -keywords+="…" -subject+="…" $filename

Stuffing the Violet tag back in
Stuffing the Violet tag back in

This shows that the “Violet” tag is now back in the current version of the file. After restarting Shotwell and doing a free text search for “Violet”, this photo now shows up whereas before it did not. It still did not show up when I clicked on “Pets/Violet” in the tag hierarchy however. It was then that I realised I also needed to put “Pets/Violet” into the “TagsList” key.

I ended up using a script to do this in bulk fashion, but individually I think you should be able to do this like:

$ exiftool -keywords+=Violet -subject+=Violet -TagsList+=Pets/Violet

After restarting Shotwell I was able to click on the “Pets/Violet” tag and see this photo.

Fixing All the Photos? ^

My process to recover from this, then, was to compile a list of each file that had been modified at the suspected time of disaster, and for each:

  1. Read the list of tags from “Keywords
  2. Read the list of tags from “Subject
  3. De-duplicate them and store them as $keywords
  4. Read the list of tags from “TagsList” and store them as $tagslist
  5. Stuff $keywords back into both “Subject” and “Keywords” of the current version of the file

Gulp.

Which files were tampered with? ^

It was relatively easy to work out which files had been screwed with, because thankfully I didn’t make any other photo modifications on 7 December 2019. So any photo that got modified that day was probably a candidate.

I haven’t mentioned what actually caused this problem yet. I don’t know exactly. At 16:53 on 7 December 2019 I was importing some photos into Shotwell, and I do seem to recall it crashed at some point, either while I was doing that or shortly after.

The photos from that import and all others afterwards had retained their tags correctly, but many that existed prior to that time seemed to be missing some or all tags. I have no idea why such a crash would cause Shotwell to do that but that must have been what did it.

Running this against my backups identified 3,721 files that had been modified on 7 December 2019:

$ cd weekly.2/specialbrew.21tc.bitfolk.com/srv/tank/Photos/Andy
$ find . -type f \
  -newermt "2019-12-07 00:00:00" \! \
  -newermt "2019-12-07 23:59:59" > ~/busted.txt

The next thing I did was to check that each of these file paths still exist in the current photo store and in the known-good backups (weekly.3).

Extract tags from known-good copies ^

Next up, I wrote a script which:

  1. Goes to the known-good copies of the files
  2. Extracts the Subject and Keywords and deduplicates them
  3. Extracts the TagsList
  4. Writes it all into a hash
  5. Dumps that out as a YAML file

All scripts mentioned here script use the Perl module Image::ExifTool which is part of the exiftool package.

backup_host$ ./gather_tags.pl < ~/busted.txt > ~/tags.yaml

tags.yaml looks a bit like this:

---
2011/01/16/16012011163.jpg:
  keywords:
  - Hatter
  - Pets
  tagslist:
  - Pets
  - Pets/Hatter
[]
2019/11/29/20191129_095218~2.jpg:
  keywords:
  - Bedfont Lakes
  - Feltham
  - London
  - Mandy
  - Pets
  - Places
  tagslist:
  - Pets
  - Pets/Mandy
  - Places
  - Places/London
  - Places/London/Feltham
  - Places/London/Feltham/Bedfont Lakes

Stuff tags back into current versions of photos ^

After transferring tags.yaml back to my home fileserver it was time to use it to stuff the tags back into the files that had lost them.

One thing to note while doing this is that if you just add a tag, it adds it even if the same tag already exists, leading to duplicates. I thought it best to first delete the tag and then add it again so that there would only be one instance of each one.

I called that one fix_tags.pl.

$ ./fix_tags.pl tags.yaml

Profit! Or, only slight loss, I guess ^

16m53s of runtime later, it had completed its work… 🙌 2020 will definitely be the year of Linux on the desktop¹.

¹ As long as you know how to manipulate EXIF tags from a programming language and have a functioning backup system and even then don’t mind losing some stuff

Losing some stuff…? ^

Unfortunately there were some things I couldn’t restore. It was at this point that I discovered that Shotwell does not ever put tags into video files (even though they do support EXIF tags…)

That means that the only record of the tags on a video file is in Shotwell’s own database, which I did not back up as I didn’t think I needed to.

Getting Tags Out of Shotwell ^

I am now backing that up, but should this sort of thing happen in the future I’d need to know how to manipulate the tags for videos in Shotwell’s database.

Shotwell’s database is an SQLite file that’s normally at $HOME/.local/share/shotwell/data/photo.db. I’m fairly familiar with SQLite so I had a poke around, but couldn’t immediately see how these tags were stored. I had to ask on the Shotwell mailing list.

Here’s how Shotwell does it. There’s a table called TagTable which stores the name of each tag and a comma-separated list of every photo/video which matches it:

sqlite> .schema TagTable 
CREATE TABLE TagTable (id INTEGER PRIMARY KEY, name TEXT UNIQUE NOT NULL, photo_id_list TEXT, time_created INTEGER);

The photo_id_list column holds the comma-separated list. Each item in the list is of the form:

  1. “thumb” or “video-” depending on whether the item is a photo or a video
  2. 16 hex digits, zero padded, which is the ID value from the PhotosTable or VideosTable for that item
  3. a comma

Full example of extracting tags for the video file 2019/12/31/20191231_121604.mp4:

$ sqlite3 /home/andy/.local/share/shotwell/DATA/photo.db
SQLite version 3.22.0 2018-01-22 18:45:57
Enter ".help" FOR usage hints.
sqlite> SELECT id
        FROM VideoTable
        WHERE filename LIKE '%20191231%';
553
sqlite> SELECT printf("%016x", 553);
0000000000000229
sqlite> SELECT name
        FROM TagTable
        WHERE photo_id_list LIKE '%video-0000000000000229,%';
/Places
/Places/London
/Places/London/Feltham
/Pets
/Places/London/Feltham/Bedfont Lakes
/Pets/Marge
/Pets/Mandy

If that is not completely clear:

  • The ID for that video file is 553
  • 553 in hexadecial is 229
  • Pad that to 16 digits, add “video-” at the front and “.” at the end (even the last item in the list has a comma at the end)
  • Search for that string in photo_id_list
  • If a row matches then the name column is a tag that is attached to that file

I don’t exactly know how I would have identified which videos got messed with, but at least I would have had both versions of the database to compare, and I now know how I would do the comparison.

Should Tags Even Be In Photos? ^

During my Twitter thread it was suggested to me that tags should not be stored in photos, but only in the photo cataloging software, where they can be backed up along with everything else.

I disagree with this for several reasons:

  • Exif exists for the purpose of storing tags like this.

  • When I move my photos from one piece of software to another I want it to be able to read the tags. I don’t want to have to input them all over again. That would be unimaginably tedious.

    When I moved from F-Spot to Shotwell the fact that the tags were in the files saved me countless hours of work. It just worked on import.

    If there wasn’t a dedicated importer feature then it would be so much work that really the only way to do it would be to extract the tags from the database and insert them again programmatically, which is basically admitting that to change software you need to be an expert. That really isn’t how this should work.

  • If the only copy of my tags is in the internal database of a unique piece of cataloging software, then I have to become an expert on the internal data store of that piece of software. I don’t want to have to do that.

    I’ve been forced to do that here for Shotwell because of a deficiency of Shotwell in not storing video tags in the files. But if we’re only talking about photos then I could have avoided it, and could also avoid having to be an expert on every future piece of cataloging software.

  • Even if I’m not moving to a different cataloging solution, lots of software understands Exif and it’s useful to be able to query those things from other software.

    I regard it very much like artist, album, author, genre etc tags in the metadata of digital music and ebooks, all of which are in the files; you would not expect to have to reconstruct these out of the database of some other bit of software every time you wanted to use them elsewhere.

It was a mistake not to backup the Shotwell database though; I thought I did not need it as I thought all tags were being stored in files, and tags were the only things I cared about. As it happened, tags were not being stored in video files and tags for video files only exist in Shotwell’s database.

Other Thoughts ^

Having backups was obviously a lifesaver here. It took me ~3 weeks to notice.

Being able to manipulate them like a regular filesystem made things a lot more convenient, so that’s a property I will want to keep in whatever future backup arrangements I have.

I might very well switch to different photo management software now, assuming I could find any that I prefer, but all software has bugs. Whatever I switch to I would have to ensure that I knew how to extract the tags from that as well, if it doesn’t store them in the files.

I don’t want to store my photos and videos “in the cloud” but it is a shortcoming of Shotwell that I can basically only use it from my desktop at home. Its database does not support multiple or remote access. I wonder if there is some web-based thing that can just read (and cache) the tags out of the files, build dynamic galleries and allow arbitrary searches on them…

Shotwell’s database schema and its use of 16 hexadecimal digits (nibbles?) means I can only store a maximum of 18,446,744,073,709,551,615 (1.844674407×10¹⁹ -1) photos or videos of dogs. Arbitrary limits suck so much.

Greyhounds Marge, Janti and Will at Sainsbury's Staines with Wimbledon Greyhound Welfare, December 2019
Marge, Janti and Will at Sainsbury’s Staines with Wimbledon Greyhound Welfare, December 2019

Getting LWP to use a newer OpenSSL

Something broke ^

Today I had a look at a customer’s problem. They had a Perl application that connects to a third party API, and as of sometime today it had started failing to connect, although the remote site API still seemed to be responding in general.

The particular Perl module for this service (doesn’t really matter what it was) wasn’t being very verbose about what was going on. It simply said:

Failed to POST to https://api.example.com/api/v1/message.json

I started by writing a small test program using LWP::UserAgent to do a POST to the same URI, and this time I saw:

500 Can’t connect to api.example.com:443 (SSL connect attempt failed with unknown errorerror:14094410:SSL routines:ssl3_read_bytes:sslv3 alert handshake failure)

So, it’s failing to do a TLS handshake. But this was working yesterday. Has anything changed? Yes, the remote service was under a denial of service attack today and they’ve just moved it behind a CDN. TLS connections are now being terminated by the CDN, not the service’s own backend.

And oh dear, the customer’s host is Debian squeeze (!) which comes with OpenSSL 0.9.8. This is badly out of date. Neither the OS nor the OpenSSL version is supported for security any more. It needs to be upgraded.

Unfortunately I am told that upgrading the OS is not an option at this time. So can we update Perl?

Well yes, we could build our own Perl reasonably easily. The underlying issue is OpenSSL, though. So it would be an upgrade of:

  • OpenSSL
  • Perl
  • Net::SSLeay
  • IO::Socket::SSL
  • LWP, as the app’s HTTP client is using that

It’s not actually that bad though. In fact you do not need to build a whole new Perl, you only need to build OpenSSL, Net::SSLeay and IO::Socket::SSL and then tell Perl (and the system’s LWP) to use the new versions of those.

Of course, everything else on the system still uses a dangerously old OpenSSL, so this is not really a long term way to avoid upgrading the operating system.

Building OpenSSL ^

After downloading and unpacking the latest stable release of OpenSSL, the sequence of commands for building, testing and installing it look like this:

$ ./config --prefix=/opt/openssl \
           --openssldir=/opt/openssl \
           -Wl,-rpath,'$(LIBRPATH)'
$ make
$ make test
$ sudo make install

The rpath thing is so that the binaries will find the libraries in the alternate path. If you were instead going to add the library path to the system’s ld.so.conf then you wouldn’t have to have that bit, but I wanted this to be self-contained.

When I did this the first time, all the tests failed and at the install step it said:

ar: /opt/openssl/lib/libcrypto.so: File format not recognized

This turned out to be because the system’s Text::Template Perl module was too old. Version 1.46 or above is required, and squeeze has 1.45.

Installing a newer Text::Template ^

So, before I could even build OpenSSL I needed to install a newer Text::Template. Cpanminus to the rescue.

$ sudo mkdir /opt/perl
$ cd /opt/perl
$ sudo cpanm --local-lib=./cpanm Text::Template

That resulted in me having a newer Text::Template in /opt/perl/cpanm/lib/perl5/. So to make sure every future invocation of Perl used that:

$ export PERL5LIB=/opt/perl/cpanm/lib/perl5/
$ perl -e 'use Text::Template; print $Text::Template::VERSION,"\n";'
1.58

Repeating the OpenSSL build steps from above then resulted in an OpenSSL install in /opt/openssl that passed all its own tests.

Installing newer Net::SSLeay and IO::Socket::SSL ^

Cpanminus once again comes to the rescue, with a twist:

$ cd /opt/perl
$ OPENSSL_PREFIX=/opt/openssl cpanm --local-lib=./cpanm Net::SSLeay IO::Socket::SSL

The OPENSSL_PREFIX is part of Net::SSLeay’s build instructions, and then IO::Socket::SSL uses that as well.

Using the result ^

Ultimately the customer’s Perl application needed to be told to use these new modules. This could be done with either the PERL5LIB environment variable or else by putting:

use lib '/opt/perl/cpanm/lib/perl5';

At the top of the main script.

The application was then once more able to talk TLS to the CDN and it all worked again.

Other recommendations ^

The customer could maybe consider putting the application into a container on a new install of the operating system.

That way, the kernel and whole of the OS would be modern and supported, but just this application would be running with a terribly outdated userland. Over time, more of the bits inside the container could be moved out to the modern host (or another container), avoiding having to do everything at once.

“Designer” dog breeds are not mongrels. They’re much worse.

It doesn’t sit well with me when owners of long-recognised breed dogs mock the new “designer” breeds as being mongrels.

“That’s not a cockapoo, it’s just a mongrel!”

Surely all currently recognised breeds were once mongrels until a human being created them and convinced enough other humans to agree that this is a good set of characteristics?

What are designer breeds made from? Dogs with certain characteristics. Who decides they are an official breed or not? Some kennel club.

All of the existing, recognised older breeds were bred for a purpose or function that primarily served human beings and only secondarily served the welfare of the animal itself. That is, these supposedly superior recognised breeds would never have existed if left to natural selection alone because they all have negatives, it’s just a question of degree. And those breeds are kept in existence again by human desires, not through any higher power. A “purebred” dog exists because humans wanted it to, just as a designer breed dog exists because humans wanted it to. Who cares what it’s called?

If someone shows me their new designer breed puppy and wants to call it a cockapoo because it was bred from a cocker spaniel and a poodle, I don’t care. Am I supposed to respect a kennel club and say “that’s not a real thing”? No, it is a real thing, it’s right in front of me.

Don’t get me wrong, I’ll be sad, because it’s yet another puppy farmed for human pleasure with no regard to its health or future, when there’s tens of thousands of healthy dogs that no one wants already in rescue kennels. Doubly so if the pairing has created something that health-wise is going to have a really hard life. But it’s not a mongrel. In some cases it could dream of being a mongrel because it would have a better life.

The name of the breed isn’t the issue. There are certainly some kennel club recognised breeds that from a welfare point of view just shouldn’t exist as their whole life is a torment, the way certain human-desired traits have been so excessively selected for. A lot of times when the “mongrel” rebuke is brought out, I think it is just something intended to annoy the owners and to convey the idea that a dog should have been rescued, not a puppy farmed. In those cases maybe a new word is needed but I don’t know what that word would be.

I just don’t get the designer breed snobbery though. When it is purely a breed issue or when it’s the owner of one type of established breed owned from birth disparaging another owner of a farmed puppy I feel like it’s really hypocritical. It’s not established vs designer, it’s rescued vs farmed.