There's a lady who forwards me e-mail all the time. I don't know who she is, but from the context I'm guessing she's 75yrs-old, give a take a decade, and lives in or about central Kentucky.
I've sent her more than one mail telling her my e-mail address doesn't belong to whoever she thinks it does, but she never replies, and the mails just keep on coming. I'm guessing she forwards things to everyone in her address book, and doesn't know how to delete or change entries in there.
I really wish she'd stop it. My life is not improved by her malice in the mail, whether she authored it or merely passed it on, blindly or not (I'm not sure which would be worse).
Perhaps an example of the flavor is called for:
> > > The American Dream ended (on November 6th) in Ohio . The second term > > > of Barack Obama will be the final nail in the coffin for the legacy of > > > the white Christian males who discovered, explored, pioneered, settled > > > and developed the greatest Republic in the history of mankind. > > > A coalition of Blacks, Latinos, Feminists, Gays, Government Workers, > > > Union Members, Environmental Extremists, the Media, Hollywood , > > > uninformed young people, the "forever needy," the chronically > > > unemployed, illegal aliens and other "fellow travelers" have ended > > > Norman Rockwell's America. > > > The Cocker Spaniel is off the front porch. The Pit Bull is in the > > > back yard. > > > The American Constitution has been replaced with Saul Alinsky's "Rules > > > for Radicals" and Chicago shyster, David Axelrod, along with > > > international Socialist George Soros will be pulling the strings on > > > their beige puppet to bring us Act 2 of the New World Order. > > > Our side ran two candidates who couldn't even win their own home > > > states, and the circus fattster Chris Christie helped Obama over the > > > top with a glowing "post Sandy " tribute that elevated the > > > "Commander-in-Chief" to Mother Teresa status. > > > People like me are completely politically irrelevant, and I will never > > > again comment on or concern myself with the aforementioned coalition > > > which has surrendered our culture, our heritage and our traditions > > > without a shot being fired. > > > You will never again out vote these people. It will take individual > > > acts of defiance and massive displays of civil disobedience to get > > > back the rights we have allowed them to take away. It will take > > > Zealots, not moderates – not reach-across-the-aisle RINOs to right > > > this ship and restore our beloved country to its former status. > > > Those who come after us will have to risk their lives, their fortunes > > > and their sacred honor to bring back the Republic that this generation > > > has timidly frittered away due to "white guilt" and political > > > correctness. I'm done.
Whatever your politics, that's hate mail, and however you couch it, I find it very difficult to stomach the idea that that passes as acceptable communication.
My thought, that I just had to say, and prompted this post: I wish it were practical to run my own MX still. Then maybe I could add rules to bounce e-mail from certain e-mail addresses with a custom message for each:
From: <MAILER-DAEMON@rivenstone.net> Subject: Undeliverable mail Your mail to <firstname.lastname@example.org> could not be delivered. I'm sorry it didn't work out. The following messages may point to the source of the problem: YOU ARE A GODDAMNED DOUCHBAG STOP SENDING ME MAIL YOU WRINKLED SCALY TWAT.
See? That's the promise of living better through technology.
I took Latin in High School. The school district had only a single Latin teacher, who had to commute between the schools daily to teach all his classes. He was an eccentric man, and many other students didn't like him; "he doesn't know how to teach", they always said. He taught me more than any other teacher I ever had. I even learned a little Latin.
The real subject of his classes was how to think for yourself and teach yourself, "how to be a life-long learner", in his words. To high school students used to being spoon-fed their lessons and trained to focus on their relationship with their teachers and negotiate their grades, he must have seemed insane.
He didn't even keep a grade book; he expected you to keep all your tests and homework after he graded them, then on the final day of the grading quarter, you were to calculate your grade and bring it to him at his desk, where he would check your math and spot-check your graded papers for honesty. His spot-checks were, erm, spotty, perhaps because teachers are graded on their student's grades. I won't rant about grade-inflation and regulatory capture just now.
Lemme bring a concrete example. He once took an entire 40-minute class period to talk about the Latin word Limen
Sidebar:It's a common misconception that English is a Romance language. English is a Germanic language with strong Romance influences through French (the language of William The Conqueror) and later, directly from Latin. This is where I get to pull out James Nicoll's quote: "The problem with defending the purity of the English language is that English is about as pure as a cribhouse whore. We don't just borrow words; on occasion, English has pursued other languages down alleyways to beat them unconscious and rifle their pockets for new vocabulary."
This is the reason why we have the duality of culinary words, like cow and beef, sheep and mutton, and chicken and poultry. The words for basic concepts like up and down and water and sky come from ancient German, as do the words for farm animals. The culinary words for the different meats come from French.
A better example might come from two Nazi-era words: "Unterseeboot" and "Untermensh". The first word is the German word that gets abbreviated as "U-boat" in English historiography. Imagine a German who speaks English as a second language pronouncing "Under sea boat" and the meaning of the first word becomes obvious. The root words for "under", "sea", and "boat" are basic words English shares in common with modern German.
"Untermensh" also shares root words with English, "under" + "man". But English tends to turn to its Romance side when discussing abstract concepts, so we use the Latin root words to translate "Untermensh": "sub" + "human".
Sidebar over, deep breath sigh. So, Limen. It has to do with the harvesting and rendering of grain, which I know nothing about. I do know the word "thresher", which is a thing that deals with "thresh". Once your have thresh, you put it in a building to keep it safe, and since you don't want that valuable stuff leaking out under the door, you install a stone beneath the door to close the gap and prevent a trail from being worn there that will expose the thresh to the outside. That stone is a "thresh-hold". That comes from ancient Germanic root words, and it has come to figuratively mean a boundry between one thing and another; by means of metaphor, the word "threshold" has come to represent the door and the doorway.
Limen is the word for "theshold" in Latin, and it's spawned many English words by the same metaphor. A "preliminary" event is something which happens before the threshold of whatever. The subliminal mind is that which underlies the threshold of consciousness.
It's metaphorical shit like this that makes language so hard for computers to translate, and also IMHO makes artificial intelligeligence all but impossible. Human brains are wired to make these leaps in non-logical ways; it's a survival instinct, maybe. Even seen a face in a pattern of leaves and branches or the bumps of a tree trunk at dusk, and had your adrenal glands dump a fuckton of nitro-whatever in your blood and make you go all fight-or-flight? It's a survival instinct; hidden faces are indicative of predators. That instinct to recognize faces in the brush is what gives us the tendency to ascribe faces to the grilles and headlights and bumpers of automobiles. Just try to convince a man's man to buy a Mazda, with it's goofy grin on the nose.
Maybe I should look up Mr. Weimer sometime.
Ooh, an external journal. I need to think about that. It seems that Ted T'so has stated that the loss of an extx external journal due to disk failure is "mostly harmless". The journal would have to go on the same disk as /boot and swap, but I don't expect to be swapping much with 4GB of RAM, and by the time that changes, I hope to pick up another 8GB or so for $50.
It still leaves the question of whether an external journal on a single (mostly) dedicated HDD is better than striping an internal journal across multiple, newer and faster HDDs, even taking into account the seeks and internal journal will create. I'm thinking an external journal will be better, due to the need to rewrite at least an entire RAID stripe for each journal update? I don't expect calculating the parity will be a significant CPU load; I will have plenty to spare.
The most relevant info on the Web I have seen talks about HW RAID, partitioned, with the external journal on an SSD, in journal=data mode. My setup would have the external journal on the slowest disk in the system, rather than the fastest, so I'm not sure that's a good idea -- wouldn't journal=data limit my write performance to that of the disk the journal lives on? How's that going to interact with the I/O scheduler when using softRAID?
I think I going to have to benchmark to answer these questions. I've toyed with external extx journals before, but was never able to convert between internal and external journals, though the last article linked above claims to show how its done. I'm not keen to try it out for the first time on a filesystem that has the only copy of my /home on it, though -- and I can't get accurate benchmarks until I wipe the 160GB disk -- but I guess I can try the conversion process itself on a mockup in a VM or something.
I also STR that akpm stated once that journaled extx FSen pin down an amount of kernel memory equal to the size of their journal, which would leave me the choice between pinning down a large amount of RAM or concentrating most disk activity (and wear) onto a tiny bit of that 160GB disk. I'm probably wrong. Crap.
Anyway, since I'm pretty sure my new Deskstars don't use the new 4KB sectors, it seems I only need to make sure my partitions are aligned on a multiple of the RAID stripe width, and pass the stride=x param to mkfs.ext4 as I have always done. I still need to learn the ins and outs of logical and physical disk geometries to work out how to actually do that, I think? Do linux softRAID superblocks offset the beginning of the first stripe, and thus all that follow? Crap.
My research, so far, is creating more questions than answers. Once I get this all sorted out, I should write about it. Yeah, right.
I ordered the parts to rebuild my main desktop (nineveh) from Newegg last night -- and when I say "rebuild" I mean "replace every part but one and reinstall". I'm only transferring one HDD to "new nineveh", but I'm also keeping my /home, which is what makes it nineveh. I'm excited. I ordered:
Due to finances, I'm going to have to wait a few weeks to buy the last bits:
Well, the exciting bit, to me at least, is what the new hardware will allow me to do. You have to understand, I've lived with nineveh, in it's current incarnation, for seven years now. Hardware has come and gone, and changes have been made, but the whole setup hasn't really changed. Now I'm making a fresh start, with a mind to eliminate all the annoyances of this configuration, 'cause I intend the core of this system to last another seven years. Really, it's the whole disk thing:
Nineveh's /home is currently spread across five partitions on a motley flock of five HDDs, no two alike; the partitions are concatenated together as an LVM volume with no redundancy. Some of those platters have no doubt spun more miles than the distance from the Earth to the Moon. Only one of the the disks doesn't have another partition on it dedicated to another task.
New nineveh will have / on RAID5, on three dedicated, identical 1TB disks. The only separate partition will be /boot (on a fourth disk), which has its downsides, but I want *all* my files to benefit from the throughput advantages of striping (for reads, at least) and there doesn't seem any good way to do that and separate /home from / without wasting money, or diskspace; or by partitioning the disks to create two RAID5 arrays on three disks -- which seems dumb.
I thought about creating the RAID array on the raw, unpartitioned disks, but I've decided to go GPT. This will probably cause me headaches down the road, but I'm too full of youthful new-hardware excitement to care.
When I first wrestled with GPT on my MacBook, I hate hate hated that there was no decent GPT equivalent to good old Linux fdisk. There is now, and I am certainly going to give that man at least ten dollars. There's even a GPT partition ID for Linux RAID now!
The single SATA drive currently in my possession will hold new nineveh's /boot and swap partitions. At first, I thought that drive would have to have MBR partitions for BIOS compatibility, but it seems GPT is compatible with legacy BIOSes as long as the bootloader is GPT-aware. I guess Ubuntu is shipping GRUB2 by default now, though, which supports GPT natively. So the "little" 160GB will have a single GPT "BIOS Boot Partition", followed by a /boot partition, and then a great big honkin' swap partition. Part of the reason for the big swap (besides putting it on a disk that won't be used for much else post-boot) is 'cause I want to be able to make /tmp a tmpfs filesystem so huge I'll never fill it.
But some research is in order:
What about partition alignment of a RAID array on GPT disks? I've always followed the instructions about creating extx filesystems on a RAID array slavishly, without really understanding what I was doing. GPT partition tables are free of some of the constraints of MBR, and now that I've decided to tread into these waters, it seems I need to get a grip on these issues if I want to know I've got my partitions and array slices laid out optimally. And yeah, until I'm sure, it will literally keep me up at night.
Then there's Linux softRAID superblocks. Linux RAID tools seem to use v0.9 superblocks by default, which are placed at the end of a partition; unfortunately, due to the partition alignment issues above, the location of the end of an MBR partition of size "yay" is unpredictable -- unless you are a disk wizard, it's likely to be off by a few (dozen) blocks, which makes squirting RAID partitions about between disks with dd dicey business. I recall kvetching on LJ about this before.
Linux softRAID v1.0 superblocks seem to be immune to this problem, from which I have inferred that they are at the "top" of the partition -- but I don't recall how I determined that with any sort of certainty. ISTR there are also v1.1 and v1.2 superblocks, one or both of which were not recognized by debian-installer at some point in the past. So I need to know what there differences there are, what d-i supports, and (again) how all this might be affected by GPT partition tables and the partition alignment thereof.
Bugger. But it's what I want. Any help, pointer, or advice is appreciated.
My plan, ATM, is to boot "new nineveh" from an Ubuntu Live CD, with only the two new drives installed. I'll install the ubuntu gdisk package, put a single Linux RAID partition on each disk with GPT partition tables, then run mdadm to create a RAID5 array in degraded mode on the partitions, with the appropriate version of the RAID superblocks and whatever esoteric alignment options I deem appropriate. The same will go for mkfs.ext4.
Then I will transfer the current contents of "old nineveh"'s /home and / to "new nineveh"'s new array over ethernet, probably as uncompressed tarballs.
Here comes the dicey bit: now I need to remove the 160GB disk from "old nineveh" -- breaking the LVM volume nineveh's /home lives on fantastically -- install it in "new nineveh", then start the Ubuntu Lucid "alternate installer" (AKA debian-installer) in expert mode, somehow convincing it to install to the degraded RAID with its goofy superblocks on the esoteric GPT disks without blowing away the existing filesystem. I need to do this because I need to repartition and format the 160GB disk at the same time, and I don't have any spare disks to back my current /home up onto.
So, in addition to all that alignment crap, I need to come up with a plan to make this swap, with "checkpoints", wherein I determine that all is well as it can be before I go on the the next step in playing musical disks with my data. And when I say "musical disks" I mean it a bit more literally than usual, because I'm going to be one disk short when the music stops. If I do it wrong, I lose all my pr0n.
And I do this for fun.
I'm posting this here because Facebook is fascist and thinks I talk too much. The question was:
C: Is anyone good at algebra? Try this on for size...8x+3y=-22 and x=31-6y let me know if this rings a bell.... I have to plot them i have to solve for both Y & X and graph them....Me: *nods* to graph them, you have to solve for "y". To single out "y" in the first one, you subtract 3y from both sides, giving you 8x=-22+3y. Then you add -22 to both sides, so you have 8x+22=3y. Then divide the whole thing by three: y=(8x+22)/3. <-- (that might better be a fraction, like:). The second one is much the same: subtract -31 from both sides, to get x-31=-6y, then divide both sides by -6 to get y=(x-31)/-6 or:8x+22 y= ------------ 3x-31 y=------- -6
Does any of this ring a bell? It always helps me to imagine the equals sign as the fulcrum in one of those scales like Lady Justice holds up: The scales will still balance as long as you add or subtract from one side the same thing you do to the other.
I'm not sure what you're being asked to do to graph the thing, but a graph usually means plugging various things in for "x" and plotting what it does to "y" until you have a graph. Hopefully you have a graphing calculator, and can just copy the thing; don't forget to note the scale.
I'm not trying to be whatever, but I have a wee bit of doubt that you need to solve for "x", partly because the second thing already is. If I'm wrong, the first one can be singled out for x by dividing both sides by 8, so you get x=(-22 + 3y)/8.
If you'd like, I'm happy to help out with Algebra for beer. Well, I'll even bring my own; you just have to let me drink it -- I can still do math in my cups.
Said offer is good to anyone in the CMH area who isn't a total cocksmoker. I can usually drive home too (there's a limit to how much you can drink and math), but failing that I reserve the right to pass out on your living room floor. Ain't nothin in this world for free. Me? I work for beer.
Corrections and/or comments to the effect that I fail at math are most welcome.