Systems

In which I eventually get round to talking about how to convert containerised applications from something like Docker into the equivalent running as LXC containers under Proxmox.

Alongside “don’t drink coffee too late at night” I have a new rule to make sure I get a good night’s sleep: “never fire up Wireshark after 10pm”.

This article documents a particularly niche issue I ran into which took a while to debug and resolve. I’ve posted it for the benefit of search engines, so that the next person to run into these exact conditions might save a few hours. You’re welcome.

This is probably interesting only to someone in my exact situation, so most people should find some cat pictures to look at instead. If you’ve been directed here by a search engine because you’re tearing your hair out, though, read on.

In macOS 10.15 Catalina, Apple have decided to change the default shell from an antediluvian version of bash to a very recent version of zsh.

One reason for this appears to be bash switching to GPL V3 a while back. Whatever the reason, though, they are flagging this change up pretty heavily every time you open a shell:

The default interactive shell is now zsh.
To update your account to use zsh, please run `chsh -s /bin/zsh`.
For more details, please visit https://support.apple.com/kb/HT208050.

Switching to a new shell will be something of a major project, so although I will probably do that at some point I was delighted to find that you can disable the warning by adding the following to your .bash_profile:

export BASH_SILENCE_DEPRECATION_WARNING=1

I am pretty paranoid about data loss. Locally, my Macs all use Time Machine, the servers all have RAID of some kind, and virtual machines are regularly backed up using Bacula. Local backup is not enough, though, so most of this is also backed up to the cloud. Unfortunately, the cloud service I’ve been using for the last few years has just been discontinued, so it’s time to pick something new.

Many of us, particularly if we have been programmers, have got into the habit of regarding computers as flawless execution engines. People with more of an electronics background tend to be a bit more sceptical, I think.

I’ve been trying to figure out why I couldn’t burn a Fedora 11 DVD to upgrade one of my oldest machines for several months now. I had checked the SHA-256 hash of the download then copied the file from the server where I run BitTorrent across to a desktop machine’s external hard drive. The burned disk verified against the image on the machine that created it but the installation self-test always failed, claiming the disk was corrupt. I tried burning from the same image on another machine; I tried burning at different speeds; I tried different blank DVDs. No change.

Finally, today, I thought to try verifying the hash on the copied image rather than the original one. It was different. Comparing the original download with the copy, I discovered two locations in the copy where byte 0x12 of a block had dropped the 0x08 bit.

It’s probably not a coincidence that the machine on which I made the corrupted copy has recently come back from a couple of extended “warranty repair” holidays during which first the main system logic board and then (at my strong and repeated insistence) the actual DRAM were replaced. The machine had been having some intermittent problems involving applications shutting down unexpectedly; these looked like memory issues to me but the manufacturer’s diagnostics had always given it a clean bill of health. As an old-school computer guy, of course, I know that the manufacturer’s diagnostics never detect real memory issues.

The moral of the story? I’m not sure there is one: “faulty hardware sometimes gives the wrong answer” seems rather an obvious thing to say. On the other hand, if you are aware of the concept of metastability in electronics, you know that there’s no such thing as perfect hardware as long as the logic needs to talk to the outside world. So we can reduce the frequency of odd weirdness to the point where we never expect to encounter it, but we can never make it go away altogether.

I’ve been using Apple’s Mighty Mouse on my desktop machines for a couple of years now. I quite like them, although the mouse’s inability to represent both mouse buttons being held down at the same time makes it necessary to keep a conventional mouse around for things like gaming.

This is a nice mouse to use, though. For example, it makes a nice solid mechanical click when you use the left or right buttons (even though there is really only one mechanical button — the whole mouse — touch sensors inside give you two “logical” buttons).

There’s even a tiny clicking sound when you squeeze the side buttons or roll the little trackball around. You can hardly hear these sounds in a normal office, but they make all the difference to the “feel” of the device. And, until today, I would have meant that literally: I’d have sworn that I could feel the little clicks through my fingertips.

Today, quite by accident, I discovered that the mouse does not make these tinier sounds if it isn’t plugged in… or, in the case of the wireless version, if you take the battery out.

Yes, there’s a tiny speaker inside, whose only purpose is to make sounds that are almost — but not quite — too quiet to hear.

This post follows on from my earlier one on Moving a FC5 Root Partition to LVM. In this episode, I conclude the work by moving the /boot partition to the new drive and do everything else necessary to remove the old drive.

One of my servers is pretty old, dating back to the dawn of recorded history — or Red Hat 7 as we now call it. Most of its storage has been migrated onto a new hard disk and the Logical Volume Manager some time ago. Finally removing the original hard drive requires a few tricky operations: moving the root partition, moving the boot partition, and moving the boot loader.

These operations are turning out to be tricky enough to make it worth writing down the procedure in case I ever need to use it again. So, here’s an aide-mémoire on moving a Fedora Core 5 root partition onto LVM.

Not installing security updates isn’t really a viable strategy these days. Even waiting a few days to see whether other people have trouble with the update is problematic when a zero day exploit might be available.

It’s a bit like playing Russian Roulette in a room full of people who feel their job is to point their guns at you until you pull the trigger.

Obviously this goes wrong once in a while. The recent Samba 3.0.23 update broke access from Windows and Mac machines on my Fedora Core 4 system, but some people with Fedora Core 5 are reporting that all logins to their systems are disabled.

After a bit of searching around and trying various things, I found that in my case I could bring my system back to life by “upgrading” to an older version of the four packages in question.

There is some indication that version 3.0.23a will be out real soon now… but that doesn’t really make me feel completely happy. Nor does the realisation that my FC4 system will officially be “legacy” next week and I’ll need an upgrade to at least FC5 to stay within my “properly supported” comfort zone.

This kind of thing does seem to happen more often with Fedora, and anecdotally seems to be related to their strategy of pulling in new releases rather than back-porting security fixes. Moving to a more “enterprise” style system for the places where I need stability rather than the latest features is probably the right answer for me; once RHEL 5 is out I will probably take a close look at it and the equivalent CentOS release.

[Update 20060729: the 3.0.23a release doesn’t fix the problem, at least for me.]

I have just spent most of the day trying to get a large SATA disk drive working in my main Linux server. Some combination of Fedora Core 3, Linux 2.6.11, the Seagate ST32508823AS drive, the drive firmware (3.1), the SATA controller, my motherboard, my motherboard settings, my processor, my power supply and for all I know my haircut and the phase of the moon are conspiring to make this not work. Well, to be precise it works really well until I try and move a few tens of GB of data on to it: then, it stops working.

I feel like a dog wearing a collar that gives it a shock when it crosses the invisible fence line placed by its owner. Something huge and inscrutable has decreed: thus far, and no further; bad dog, don’t go there. I could spend another day and perhaps make some progress, or perhaps not. Experience has taught me that the odds aren’t good that I will ever really understand the problem.

So instead I will buy an equivalent IDE drive for less money than I could earn in the time I’ve already wasted, and it will work and the collar will stop giving me shocks. Perhaps one day I will find a system I can make this drive work in.

This is not the most satisfactory outcome, but all I can do about it is bark in frustration.

I’ve been thinking of replacing my now geriatric Hewlett Packard laptop for some time, and the idea of getting a nice Unix laptop has been getting more and more attractive.

So I spent an hour in the local Apple dealer today picking out some hardware. “I’ll take one of those”, finally.

“Oh, we don’t actually have any in stock, sir. Besides, you wouldn’t want to buy one today of all days.” …and the salesguy looked at his watch.

I guess everyone except me knew that Steve Jobs was giving his famed New Things keynote at MacWorld in a few minutes. D’oh.

[Updated: turns out Apple didn’t update the Powerbook line after all. The new Mac mini is pretty cute, though, and worth looking at for a small headless Unix box.]

I’ve long held that the shift from unquestioning youthful enthusiasm for whatever operating system you grew up with to general disillusionment with all of them (except for ones you’ve written parts of, naturally) is an essential rite of passage. In short, Every OS Sucks and it is just a question of which one is least bad for whatever you want to do right now.

Today, I found that some years back Three Dead Trolls In a Baggie recorded a grumpy song celebrating this grumpy attitude, and you can download Every OS Sucks for free from their Ampcast page. If you’re bandwidth rich, they have a video version too.

[Found through Adam Curry’s audio show.]

[2018-02-20: Removed dead Ampcast link. Video also available on YouTube]