Hacker's Diary

A rough account of what I did with Emacs recently.

November 27

Synology update: after spending a few days shuffling things around and discovering drives that were unhealthier than I thought, I ordered a pair of Seagate "IronWolf" (seriously?) 4TB drives. In the meantime I've continued sorting through the contents of the Drobo and various other disks I have, uh, lying around. Among the discoveries are two Time Machine backups of no great consequence which were somehow completely empty - presumably part of one of my grand experiments to merge backups - and I've no idea what I did with the content but, as I say, no great consequence; also two possibly identical backups which again I presume were part of some previous fiddling around. I am currently scrubbing those to ensure I can safely nuke one of them; moving a Time Machine backup to the Synology; and attempting to migrate a saved-to-local-disk Time Machine backup into a saved-to-network-drive Time Machine backup. A certain amount of the above is being overly cautious with things I maybe could just delete, but I've learned the hard way that it's all too easy to lose the one digital copy you have of a thing through carelessness.

Had a round of poking at the undead TRV with a diagnostic tool. It's not great, to be honest, as it's limited to certain generic actions. So, for example, a Z-Wave system has a controller which maintains a map of your Z-Wave network. You can't interact with that directly: node IDs are allocated by the controller and can't be modified, for example. When a node goes bad, it's supposed, in theory, to be marked as failed by the controller. So using the diagnostic tool, I tried "mark node as failed", to which the tool responded, "node is responding!". After some period of trying various things, it eventually marked the node as failed. Great! "Remove node from network", I suggested. "Remove node from network timed out", it replied. Repeatedly. At this point I gave up on graceful options and figured I'd hard-reset the TRV, then clean up later. Except... even the hard reset - a manual mechanical process on the dubious device - didn't seem to work. At which point I tossed the thing in the corner and gave up on it.

That was during the week. Today I picked it up, pressed the button, and it was no longer doing its strobing-backlight routine. Huh. Checked the Z-Wave system and it says, "hey! found a new node!". Which means that, ok, I can add this back into the system, but it now shows up as node 26 instead of node 22 and there are a few things that are manually connected to the latter which need to be found and converted to the former. Given this has happened to me several times already (I do not have 26 Z-Wave devices; the automatically-allocated ID gets incremented every time I go through this dance) I should probably set aside at least a little time to hacking up a script of some sort to clean up the mess it leaves behind it.

November 26

Red Sparrow was pretty much perfect. Knew nothing about it going in, was completely hooked throughout. An excellent spy thriller from start to finish; a little bit squeamish in spots, a little bit gratuitous in spots, but great story.

November 13

Migration process under way. It's a bit slow as I'm trying to move things around on USB-connected drives to make them palatable to non-USB-connected drives, e.g. "native" Time Machine backups. Apple helpfully suggests using the Finder to drag and drop things, but when the files are > 1TB that means it can spend literally hours days "Preparing to copy..."

November 9

Synology DS420+ in da house! Now to see about migrating stuff to it...

November 8

Wrapped up Season 2 of Elementary, wherein the Fellowship is Broken, or something. Weird to be going through these cram-it-all-into-45-minutes episodes only to get a three-parter for the finale. Anyway, roll on season 3.

November 7

New season of Doctor Who! Yay!

November 3

The rabbit holes... must check on that thing on the Mac Mini... hmm, it all seems a bit sluggish, I wonder what's wrong?... wait, it's writing to swap? Huh. Lemme fire up the activity monitor... OK, so servermgrd is consuming 5GB, and that's making eventsd thrash, so eventsd is now up to 2GB and this is probably a self-feeding cycle, so I'll kill off the Server Manager front-end and see how it goes... OK, swap usage has halved, but it's still cranking the disk a lot... Ah. Spotlight is indexing the slow disk, so that's gonna account for a whole lot of reading, writing, and thrashing. Best leave it get on with its work.

November 2

One of my z-wave TRVs has flaked out a couple of times in a sufficiently unique way that I can't find any useful online advice: it stops reporting temperatures or accepting commands, but I think it's still responding to low-level probing because as far as I can tell the controller is able to ping it and get a response and basic capability information. Touching the "wake up" button results in it rapidly flashing its backlight, and both the "network" and "alarm" icons blink constantly. It still seems to manage temperature regulation, you just can't find out what temperature it's measuring or make it hotter or cooler. Compounding this is that it's a battery-driven device, and my various readings on z-wave have led me to conclude that it will never be marked as dead - it's been sitting here on the sofa with the battery removed for more than a day and the controller insists it's alive despite not having heard from it in all that time. This makes it impossible to mark as failed, because the controller insists it's not dead, just resting, and if you can't mark it as failed, it's sort of hard to remove it from the controller. Now, the other thing I've done in the past is to factory-reset it, which means - as far as I can tell - wiping the network association including the controller-assigned ID, so when it rejoins the network it gets allocated a new ID. And then I have to go and manually update a bunch of things so that the new ID is recognised as the appropriate device etc. etc. etc. I've not yet given up hope that the controller will eventually mark it as Actually Dead, at least not until I can find definitive confirmation of my half-recalled knowledge of how it's supposed to work. One thing I will observe: it's the first TRV I bought, and it's a different brand to all the rest albeit with what should be the same internals, and the same physical appearance. So it's plausible there's some sort of glitchy firmware at play here and that's why it's only this TRV and not any of the others.

previous month | current month| next month