ddrescue has completed! I’m about to let this image to a new drive to recover the non-system files which were unreadable. But I do want to compare-and-contrast something. Here’s ddrescue’s completion output. Note in this list:

• 5.49 gigs of errors (recovered, tho’ it took four days of processing
• 21261 read errors.

About to copy 320072 MBytes from /dev/sdb to /home/kahvi/Archives/clearbrook-image.img
    Starting positions: infile = 0 B,  outfile = 0 B
    Copy block size: 128 sectors
Sector size: 512  bytes
Max retries: 0
Direct: yes    Sparse: no    Split: yes    Truncate: no

Press Ctrl-C to interrupt
Initial status (read from logfile)
rescued:   314584 MB,  errsize:   5488 MB,  errors:   21261
Current status
rescued:   314584 MB,  errsize:   5488 MB,  current rate:        0 B/s
   ipos:         0 B,   errors:   21261,    average rate:        0 B/s
   opos:         0 B,     time from last successful read:       0 s
Finished

Yikes! Clearly the drive is on its last legs! What does Hitachi’s SMART firmware think?

smartctl 5.41 2011-06-09 r3365 [i686-linux-3.2.0-39-generic-pae] (local build)
Copyright (C) 2002-11 by Bruce Allen, http://smartmontools.sourceforge.net

=== START OF READ SMART DATA SECTION ===
SMART overall-health self-assessment test result: PASSED

oh.

THANKS GYES.

Anna’s new book is out! It’s called Valor of the Healer (yes, spellcheck, because Carina Press is an American publisher, leave it) and you can download it right! now! for an eminently reasonable new-release price. And you should! Because I said so. Also because it’s good. And I helped make certain, um, possibly-seen-as-nefarious mechanisations work better.

Yes, that does in fact mean I’m a consulting supervillain. SHOVE OVER, MORIARTY, YOU AREN’T THE ONLY GAME IN TOWN. Do you have a heat ray? No? Okay then. SHIFT IT.

Clickie!

I am telling you guys, it is so much more fun being on this side of the spoiler wall. It is. XD

A heads up: today’s Tech Woot looks to be a quality microphone for the price. I didn’t even know Blue – who are a quality maker – made a USB microphone, or, for that matter, much in this price range. (I think they had a cheap ribbon mic at some point? But mostly their ribbons are pretty high end. And this is a large-can condenser, anyway.)

I wouldn’t want to move it around a lot – it’ll be too fragile to take out – but anyone wanting to muck around with home recording for as little cost as possible, just wanting to learn the concepts? You could do a lot worse. Particularly given that it’s a USB mic, and you won’t need to buy an external interface.

Here’s the URL: http://tech.woot.com/offers/yeti-usb-microphone-12

And, for the record, I have no connection or financial interest in either Woot or Blue, other than losing in a bidding war for a Blue Woodpecker ribbon mic once on eBay. XD

(h/t Stickmaker for the callout.)

Anna and I were in Victoria last week, and she picked up a mammoth at the provincial museum. (And, by the way, I had no idea how fucking big those animals were. Holy hell in a handbasket those things were big. I know, I know, they’re called mammoths, but ‘jumbo shrimp’ are called ‘jumbo’ and that’s not true now is it?)

I posted lots of pictures to Twitter about ADVENTURES! And now, mammoth needs a name. And she was debating whether mammoth was male or female and I said not to force mammoth into a false gender binary, and now we just need a name.

The post is over on her blog. Go vote to name our genderqueer mammoth! Also, respect genderqueer mammoth’s identity, or they will trample you into viscera. Just see if they don’t.

Name me!

We’re down to the last bits of reset/upgrade/recovery on all the systems; I’m finally monkeying with the Hitachi hard drive that started throwing read errors in my laptop.

SMART status check reports zero (0) read failures on this Hitachi drive ever. That’s what the drive’s firmware thinks.

The gddrescue disk recovery tool is down to reading 4144 bytes per second a it tries to chew through all the read errors on this first pass; it hasn’t successfully read a whole 512-byte disk block in 1.1 hours.

I think we’ve found our “SMART got ruined by marketing departments” champion.

eta: The Procession of the Dead:

It’s my birthday! At least, my legal one. There’s another date which is candidate for Actual Birthdate, but that’s a whole ‘nother story.

And I’m spending it moving boxes and files and hard drives. YAY. I can’t entirely believe I’m still moving files around from the techsplosion and Ubuntu upgrade farce, but I am. No, no additional hard drives have failed, we seem to be past that for the moment. But new drives mean larger drives mean the old backup drives aren’t big enough anymore means moving things around means playing Towers of Hammurabi with archives.

And y’know, moving a few hundred gigs at a time over USB2? That’s… not the fastest thing ever.

Home stretch, tho’. Home stretch. I should set up a Hall of Remembrance for all the dead drives. The Lord of All Drives could preside over it.

17 Years Good Service

Seriously, that’s the weird little what-is-this-doing-here 2G hard drive from lodestone that was serving as lodestone’s swap. I looked it up. It’s seventeen years old. And in fine working order! I figure that makes it RULER OF ALL HARD DRIVES, because damn.

And on a not-entirely-dissimilar note, what do I do with 160 and 40 gig EIDE hard drives in good working order, anyway? The small one is really slow, to be honest – it’s a total dog – so it kind of sucks and I’m not unwilling to drill it, but the 160g is reasonably fast and low milage and everything! What do you do with stuff like that?

I also have three dead drives, and one 500G EDIE drive that got yanked but which I’m putting into a housing and back into service in a new capacity. (Archives.) That’s six hard drives I’ve had to pull out of machines to get this all back together.

“Digital is forever,” they say. “Once it’s online, it’s eternal.” What a load of crap. Reality? Everything is super fragile and needs constant maintenance.

Even if you’re not a supervillain.

I was going to write more about Victoria and the trip – I talked about a open mic appearance on Friday already – but my Macbook drive is more fucked up than I feared.

Anybody know how to make Time Machine skip past files with read errors and not just bail out?

See, I currently have no backup because Time Machine last week decided the backup was damaged and that it needed to start over with a new one. It’s done this before over the years so I said sure, go ahead.

It deleted the old backup and refused to make a new one. It gets about 1.7 gig in, then fails due to what it calls probably-transient network errors.

What it’s actually failing on are read errors on the laptop drive. Read errors disk utility can’t find. Read errors fsck can’t find either. Read errors SMART says aren’t happening.

But there are a whole bunch of files that if you try to copy them generate read fails. dd sees them too, and fails, if I try to image the disk. I went through and generated a massive list of bad files – and there are many – by having the system cp -pr them all indiviually while I was in Victoria.

I’ve currently added all of them to the Time Machine exclusion list and am trying to get a backup that way. I rather suspect this, too, will fail, due to a previously-undetected bad file.

Does anyone out there know whether there’s a way to make Time Machine not bomb out on these read errors? Or failing that, have another, good solution? Because I really need a better backup than the results of a big tree of cp -pr.

But at least I now have that. Before I set up that job, I didn’t even have that. Not after Time Machine’s lies.

eta:: With a bunch of exclusions added, I have a time machine backup of most of the drive. But it could be better with fewer exclusions if you know how to make Time Machine skip files with read errors instead of failing out. Can this even be done?

Apparently I have an opposite number here in Victoria – the male version of me, musically. A bunch of the audience were all, “We have GOT to get them both in the aame room!” But then others were, “No! No! That would be very danerous!”

I guess that makes me the gatekeeper and him the keymaster? I’m not sure. THERE IS NO MUSIC ONLY ZUUL.

Either way: STEWART IN VICTORIA! THERE IS A NEW SUPERVILLAIN IN TOWN AND SHE IS CALLING YOU OUT. By whom I mean mean me. I’m calling you out. See how that worked there?

Anyway that went well, my just-over-the-head-cold voice let me make it though my songs, if just barely. (Halfway through Anarchy Now my throat said, “TWO MINUTES! That’s all I got!”)

I want to embed a picture Anna took, but it’s not letting me, because ARG DIE. So here’s a link. Fernwood is great. Cornerstone Cafe is also great. Victoria: so far, you’re kind of awesome.

On the Clipper ferry heading to Victoria! I’m not as fond of the Clipper as the other ferries; it’s not bad but it… it’s kind of airplanish. Not as bad as an airplane, not at all, but there’s no lounge or anything like the train or the peninsula ferries.

My head cold seems to be letting up, so I might do an open mic at Cornerstone Cafe, but no promises. I haven’t played since Monday, between replacing hard drive – we’re up to five, honestly, what is this – and this cold, and the post-convention cleanup.

Oh, and I managed to get Ubuntu to reinstall the 3.2.0 kernel, and this time it worked! So I’ll be putting that through its paces as soon as I get home, but in initial teating, we look pretty good. Most importantly, my weird hardware is still working. The funny part is that the more modern version of Jack sees, complains about, and reports the device enumating things wrong – the problem which crashed the 2.x kernels (!) which prompted the 3.1.5 install to begin with.

All of which means basically nothing to anyone! Except that it means things should work better in general in production. And I can use other plugins I couldn’t use before, which is awesome. I’ll be downloading those on Monday. 😀

Anyway, have a good weekend, everybody! Anybody going to be at the Le Vent du Nord show tomorrow?

I’ve been taking advantage of this little schedule break after Norwescon to try to upgrade my DAW from Ubuntu 10.04 LTS to Ubuntu 12.04 LTS, because support for 10.04 LTS is going away this month, and also because there are a lot of fixes I need in later versions of Jack and Ardour, and Jack setup and building is so strange that even the author group says DO NOT TRY THIS AT HOME.

I SAID NO SILVER M&Ms IN THE MIX!

It has been an insane nightmare. Had I not been working against a new hard drive onto which I had cloned my old setup, I would’ve been brutally screwed. Did you know 12.04 just bricks some machines – like, send-it-back-to-manufacturer brick them – at startup? Did you know 12.04 upgrade can and will render your machine unbootable? (That happened to me through the GUI; I had to re-image the drive. Also? The install disc for 12.04? I never even get to the first setup dialogue. Hangs.) Did you know that if you try to do a stepwise upgrade as per the instructions here that the tool you use to do it is hardcoded to look in the wrong place for the upgrade files, and that this bug is known and supposedly fixed but still happens to me?

Shall I go on? Because I can. This is why I smashed an Ubuntu install CD yesterday out of frustration and rage. (See above.)

Anyway, I eventually got the server upgrade path to work – it was literally the last route available, but it got me there, mostly. After putting my machine in a state which would leave it unbootable, it had the decency not to force a reboot, and after a few hours, I fixed it. This is also a known bug. If you upgrade in the GUI, you’re just pooched. As I was, except I was working off a new image, my original drives untouched, so I could start over.

Even with all of the above, I’d currently be dead in the water again(!), except the 3.1.5 kernel I installed myself to work around a combination of kernel bug and ill-behaved USB external sound hardware which enumerates its own hardware incorrectly(!) is booting fine, and running fine, under 12.04. So I’m actually up and running! As of around 2am this morning.

However, the kernel the installer wants to install is 3.2.0-39, and it panics at startup. That’s a later version, and I’m worried that this might bite me in the ass somewhere.

Will it? Anybody know? Will 12.04 be stable under a 3.1.5 kernel?

3.2.0-39 doesn’t even get loaded. Here, see if you have any ideas:

Starting up...
[0.929456] Kernel panic - not syncing: VFS: Unable to mount root fs on unknown-block (0,0)
[0.929507] Pid: 1, comm: swapper/0 Not tainted 3.2.0-39-generic #62-Ubuntu
[0.929552] Call Trace:
[0.929594] [<c1561988>] fukkit I'm not typing all this in. printk, panic, mount_block_root, ? sys_mknod, mount_root, prepare_namespace, ?sys_access, kernel_init, ? start_kernel, kernel_thread_helper.

The 3.1.5 kernel (same drive, same directory, same install, etc) launches fine.

So even if I get this working completely, I’ll be looking at a new distribution next time. This is obscene. The fact that pretty much everything I ran into – once I got the drive cloned, where I hit other problems, such as grub insisting that I had no hard drives after booting from one of them and mounting three – is known, and that the core dev team is basically okay with that tells me I’m kind of done with Ubuntu.

Mageia has been recommended highly. So has Mint, but I was later told Mint is just Ubuntu with a different GUI.

Got anything to declare?

Yeah. Don’t install Ubuntu.