« first day (698 days earlier)      last day (4298 days later) » 

6:00 PM
Your core question isn't about TCP/IP though - that's just the spot you want to stick your fingers.
You're really asking "How to I go about modifying/extending the Linux kernel" -- Stack Overflow (site or chat) may have some helpful input too
 
@lsiunsuex And anyway, WordPress isn't really the problem, it's people adding shit to WordPress that's the problem. Last week I saw a WP theme with its own upload script. Which just happened to be the point of compromise for the server.
 
@Bruce have you tried asking on SO or it's chat rooms which is more the level you'll be working with
 
@MichaelHampton my problem with wordpress is not the blog part. as a blog, its fine. as you, the asshole who wants to use wordpress as a shopping cart is the one i have a problem with
 
@Iain: I did. My question was closed. I will rephrase and ask it again. Thanks @Iain @voretaq7 @JustinDearing
 
@lsiunsuex Oh yes, there are much better shopping carts.
 
6:02 PM
@Bruce Probably a better thing to ask about in chat -- on the site it's probably going to be looked at as a Shopping Question
 
i rewrote a website done in wordpress, using a shopping cart, with a cms, in strait up code (no frameworks) in 2.5 days flat (friday night, sat, sun, and cleanup on monday) - with commenting system and a basic blog
 
@lsiunsuex rails?
 
i prefer php
i'm a php whore
 
@Bruce Ultimately kernel programming comes down to "Get a copy of the kernel, dive in, and don't break shit." (that last part is hard)
 
@lsiunsuex then fix test.laspecialty.com !
 
6:03 PM
@voretaq7 When you put on your DBA Hat, what sort of queries/business logic are you working on, if you don't mind my asking
 
@ewwhite told u to contact my designer
 
@lsiunsuex Ah, I only asked because I've been staring at rails lately. It seems quite interesting.
 
DO IT!!!!!!!! :)
 
@JoelESalas "nothing interesting" unfortunately
 
@voretaq7: thanks for the tips and resources.
 
6:03 PM
i dont work with clients directly. my designer does. i work with my designer. prevents me from calling the client an asshole and getting sued :)
 
(basic "Find the things that happened in this time period so we can bill for them", "clean up these records which are malformed", etc.)
 
@voretaq7 I'm trying to get a feel for the type of benchmarking and metrics I'll need to do A-B testing of configuration and schema changes
 
@Bruce Hey! There's an echo in here!
 
@JoelESalas ohh, you mean in terms of administration, not design?
 
@ChrisS: ?
 
6:05 PM
@voretaq7 A bit of both. We have a performance conundrum and I've been asked to find the sore spots
 
@JustinDearing heck, the Windows socks implementation is the same at the API level as BSD. Only the newer API (forget it's name) is different
 
Setting the slow-query threshold, occasionally looking at the logs and seeing what shows up.
If something shows up a lot profiling it with EXPLAIN / EXPLAIN ANALYZE and figuring out where indexing might help
 
@Bruce You posted that exact same thing this morning. And we told you the same thing then.
 
@ChrisS even the new one is still BSD Sockets if you look past the window-dressing
 
@MichaelHampton rails has its uses. rails is also a framework though on top of ruby (re: hate frameworks)
 
6:06 PM
@voretaq7 I haven't dug into it, but I thought MS redid the stack top to bottom.
 
so like php = ruby , cakephp = ruby on rails
 
@JoelESalas do you know which queries are slow already? :)
 
@lsiunsuex Never touched cake.
 
neither have i - just saying :)
 
@ChrisS: I know. I was interrupted in the morning by someone and couldn't complete the conversation
 
6:07 PM
@ChrisS I know they did a major rewrite, but IIRC the primitives are all still there, just a few have different names :)
 
@Bruce So what exactly at you trying to accomplish.. I know you want to know more about IP, but to what end? That would help us point you in the right direction.
 
@lsiunsuex So, frameworks are never worth using?
 
never say never. but i prefer not to use them
 
@MichaelHampton It is better to use a well designed one, then to reinvent your own every time you build a site.
 
@ChrisS: chat.stackexchange.com/transcript/message/5711143#5711143. I have implemented the scheme in ns2 till now.
 
6:09 PM
@MichaelHampton In any project of sufficient complexity you will have to write your own framework.
For any project below that level of complexity if you can find a framework you like, that gets out of your way, I strongly advise using it.
 
@Bruce Ah, missed that one, my apologies. Congestion avoidance at what level? Ethernet, IP, TCP, Application?
 
@ChrisS: I don't need to know more about IP. Sorry for the bad phrasing of the question
@ChrisS: No problem. TCP
 
@ChrisS Isn't ethernet congestion avoidance "replace your fucking hub with a switch you IDIOT!" these days? :)
 
@lsiunsuex Interesting. I would have thought frameworks would be useful time-savers in many circumstances. What did you find was the down side?
 
@voretaq7 There still is broadcast/multicast traffic that could cause congestion.
 
6:12 PM
I can't believe a client called me at 11 am. Gaaaaagh.
 
@voretaq7 Yes, but I don't know what he's working on... Maybe he's doing a project to graft FC style flow control into Ethernet.
 
what is this networking thing all about?
just as long as it's not finance-related.
 
i've done over 40 websites in the last couple years (a lot? a little? i do it on the side either way) - every single website, although from the same designer, is always different. yeah, i'm always connecting to mysql, i'm always writing a blog system with comments and users. but their always different in some way.
 
@Zoredache oh yeah you can do broadcast on ethernet can't you... fucking masochists.
 
once i write something (blog, comment system, whatever) - i save it obviously - database and all. the next time i have to write that system, that gets copied, integrated and modified to fit the site.
thats my framework. past code being used in new systems is my time saver - still pure code, but consistent, never reinventing / wasting time
 
6:14 PM
We're replacing your HFT backbone with waterlogged thicknet. Let's see if the trading houses notice...
(STOCK CRASH GOES HERE)
 
@voretaq7 That's one thing IPv6 got right, removing broadcasts.
 
@voretaq7: Doesn't using VLAN solve the broadcast problem? I have heard some switches even drop broadcasts if a host is sending excessive amounts of them
 
I was just writing about how Chicago-New York trading lines are moving to microwave wireless...
 
@Bruce So you're already familiar with tools.ietf.org/html/rfc5681 ?
 
(relaying tower-to-tower...)
 
6:15 PM
i think of it as code snippets but on a grander scale. i'm not snipping a couple lines of code, i'm snipping entire pages / processes
 
@ChrisS: Yes I have read it
 
@lsiunsuex Yep, sounds like you wrote your own framework. So you know every line of code in it. (Except the ones you forgot after not seeing them for a year... :) )
 
@ChrisS well they really didn't remove it, they just pushed it into link-local multicast (all-nodes) where it probably belongs
 
But there's been talk of solar-powered drones with microwave antennas set to relay traffic over distance...
 
You can still ping (whatever the fuck the address is that I don't remember) and get a reply from everyone on your segment
 
6:16 PM
@voretaq7 Yeah, but at least with multicast only the hosts that want the traffic get it... Not everyone indiscriminately.
 
@MichaelHampton for sure - getting a call to fix something / change something on a website i haven't touched in a year really puts you on the spot. Why the f' did i do this that way? where did i put that config file? ... always learn new ways to do shit as time passes
 
@Bruce VLANs just segment the broadcast domains
@ChrisS there are legit uses for broadcast (pings anyway)
 
@voretaq7: Doesn't that make it manageable?
 
When I inherit a network one of the first things I do is a broadcast ping. Identifies 99% of the shit with an IP :)
 
@Bruce What OS are you hacking on?
 
6:18 PM
@Bruce it reduces the scope of the problem. If you break your arm and splint it you've reduced the scope of the problem, but the arm is still broken :-)
 
@ChrisS: Ubuntu
@voretaq7: lulz
 
Granted I'd rather have my arm in a splint (and my broadcast domains separated by vlans and subnets), but the core problem still remains :-)
 
@Bruce amazon.com/… and/or diving into the code...
 
@Bruce the book @ChrisS linked to is probably a good starting point for you -- I think it's recent enough that it would cover modular congestion control algorithms if they're currently implemented in Linux
 
@voretaq7 I know Linux had some TCP congestion control at that time, but I don't know if they had RFC 5681 fully implemented (or even if they do now....)
 
6:23 PM
@Bruce you may also want to consider doing your work on FreeBSD -- the BSD TCP/IP subsystem was the granddaddy, and FreeBSD has the modular congestion control framework so there's already a way to drop your code right in. See this paper
 
RFC 5681 wasn't published until a year after the book. But those topics were kicked around for a while before, and the drafts, etc.
 
@ChrisS: Thanks for the suggestion. RFC5681 is only slightly modified as compared to RFC2581 (1999) and RFC 2001 (1997). @voretaq7: That might be easier and I can use the awesome Stevens book for it.
 
@Bruce IDK if the Stevens book covers the modular congestion control bits, but the freebsd mailing lists can certainly point you in the right direction (or you can use one of the existing CC algorithms as a template to drop your code into)
 
@voretaq7 ok
 
That's one thing about hacking the BSD kernels, they're tightly controlled, so they've got beautiful design. It's very easy to pickup the design patters and figure out what's going on. Everything thing's got mediocre to excellent commenting too.
 
6:36 PM
@ChrisS yeah, and the TCP code is pretty clean
years of refinement and people going "What the FUCK?!?" and fixing strange things is good for that.
(Still the highest density of the ternary ? that I'm aware of though)
 
Though I looked through telnet's code the other day and it's got 483 comments in 11k lines of code.
 
@ChrisS How many of those comments are "Why are you here?"
 
@voretaq7 We do some gnarly auto-generated queries with like 50 JOINs in 'em
but most of those will be magic-wanded away by the use of an actual search server
 
@JoelESalas multiple JOINs are a bitch if they're auto-generated
 
@voretaq7 287 occurrences out of 68k lines of code.... Not that bad.
 
6:39 PM
you always want to reduce search scope as fast as you ca
@ChrisS Every time I parse that operator the little man inside my brain gets kicked in the face by a donkey
 
@voretaq7 The vast majority are documenting typedefs and structures... With really, really lousy comments. Like struct env_1st *next; /* pointer to next structure */
Sure glad they put that comment in there. I've never seen a double linked list before!
 
@ChrisS YOUR TELNET COMMENTS ARE BAD AND YOU SHOULD FEEL BAD!
 
@voretaq7 I don't think using ternary ?: is intrinsically bad, then again I format it over three lines so its readable justaprogrammer.net/2010/04/07/…
 
sports.yahoo.com/blogs/olympics-fourth-place-medal/… posting for the main photo, not the ridiculous 5 paragraphs to explain why sand doesn't stick to half naked women in london
 
@voretaq7 Exactly, I really can't optimize some of these queries without greatly impacting others
 
6:46 PM
@JustinDearing there's nothing wrong with it - I use it myself sometimes - but I don't see it often enough so my brain has to remember what it is
@JoelESalas if you EXPLAIN ANALYZE them and find the part that reduces the search space the most you can move that so its the first JOIN performed
 
I don't use ternaries. if doesn't take up enough additional space for the readability tradeoff.
 
@voretaq7 Interesting idea, I'll see if we can try that.
 
after that (and as long as all the columns you join on are indexed) you're pretty much done.
 
@voretaq7 Also, disabling journaling on OSX == It's almost usable!
 
@JoelESalas yeah but then you have a power failure :P
 
6:48 PM
@voretaq7 well I don't think I ever used it in proper G-d fearing ANSI C. C APIs tend to not lend themselves to reasonable uses of the ternary operator
 
I'd disable journaling if Safari didn't lock up my laptop
 
@voretaq7 I don't care really, anything good is backed up
 
@JustinDearing yeah most APIs and frameworks are written in a way that any use of the ternary operator should be inside. Like the TCP stack :)
 
@JoelESalas disable postgres journing on OSX or disavble filesystem journaling on an HFS+ volumn storing postgres
 
If you need to use the ternary operator you need another layer of abstraction :-)
 
6:50 PM
SO who was in on the bet about the server losing network link every so often?
 
@ewwhite Can I still place a bet
 
@ewwhite cable?
 
@lsiunsuex It's not "sand", it's silicon dioxide (which is the main component of beach sand, but typical sand has a lot of other crap in it)
 
I just got a blast email from customer...
"Today at 3 PM we will perform emergency network repairs to replace a faulty switch. For about 5 minutes you may experience a loss of connection to all servers"
 
@ChrisS In LA, that means bits of glass, cigarette butts, corpses, etc.
 
6:52 PM
this is the third faulty switch that's been replaced in 2 months...
 
@ewwhite What kind of switch? Dlink?
 
Linksys
 
@ewwhite Obviously the faults are not originating with the switch...
 
could be power...
I think there's a power issue... I've never blown through a ProLiant motherboard after a week
 
@ewwhite My first thought was to look for an idiot.
 
6:54 PM
There's one there :
 
Ha ha! Linksys. >=]
 
So for you old school UNIX people, I read today that CDE was open sourced.
 
@ewwhite UPS -- It's NOT just a shipping company!
@MichaelHampton yeah, there was a resounding "so what" a while back :)
I already have OLVWM, why the fuck would I want CDE? :)
 
@voretaq7 Haha, it brings back memories. It's like playing with the old C64.
 
@MichaelHampton No it isn't. I LIKED my C64. CDE sucked :P
 
7:00 PM
Well, there IS that...
 
File System Visualizer, also known as fsv, is a 3D file browser using OpenGL, created by Daniel Richard G. It is a clone of SGI's fsn file manager for IRIX systems, aimed to run on modern Linux and other Unix-like operating systems. It is capable of representing file systems in two ways - in MapV mode, files and directories are represented as cuboids of equal height, with the size of the cuboid representing the size of the file or directory, and in TreeV mode, files and directories are shown in a more conventional tree manner, with links between parent directories and subdirectories, and...
 
@voretaq7 The UPS was leaking...
(acid?)
@voretaq7 Some sort of high-end Eaton UPS... but it was replaced three times in 4 years.
 
@JoelESalas I saw that movie. It sucked.
 
I may be joining you NYC fools...
 
@ewwhite I was secretly hoping you'd sign up with SpaceX. That way humanity stands a chance of living on Mars this century
 
7:03 PM
@ewwhite Buy an APC? :)
 
@JoelESalas The job descriptions on Glassdoor were rough... "They wouldn't let me use the bathroom during the day! I kept getting poked by a sharp stick at my desk..."
 
Stop drilling holes in the bottom?
 
Hey, quick sanity check. (Again. This job's gonna kill what little is left of my brain.)

If your only choice for a smallish remote site's internet connection was between a T1 and "business-class" broadband, you'd go with... [_], for the painfully obvious reasons that... [_]?
 
@HopelessN00b Rank in order of importance: Bandwidth, Cost, Uptime
 
@voretaq7 I'd need an APC RT 8000VA to be equivalent to that's there. Luckily, it's not my problem... except for the machine that loses link.
@HopelessN00b I'd get the most reliable and the least expensive lines... and use a link balancer to handle the rest..
so T1 and shit ADSL... or maybe cable.
 
7:06 PM
@HopelessN00b Having had Comcrap Business for five years, and my last T1 back in the 90's... For the one real mid-day outage I had, the tech was out in 15 minutes and it was fixed in about 90 minutes.
 
sell them on the idea of being able to shape traffic and activity flow based on a variety of conditions...
 
(Right, details details.)

Bandwidth, Cost important, uptime not so much. As in the SLA for business class broadband would be more than sufficient.
 
and then that the lines can failover for each other.
There's no excuse not to have multiple internet lines these days.
 
@HopelessN00b business broadband (provides bandwidth, followed by cost, followed by uptime)
 
@HopelessN00b Broadband!
 
7:08 PM
@HopelessN00b They also occasionally do overnight maintenance around 1-2am that they conveniently forget to mention.
 
@ewwhite yes there is: "Can't get two providers"
 
@voretaq7 I have one site with DSL, an Elfiq and a 4G usb stick sitting as failover.
 
I've been trying to get Verizon in here to get a site survey for a year now.
 
@ewwhite T1s get very pricey when Cable doesn't go down for 2 years and your boss wants to know why we're spending $300/mo on the T1 that's never used.
 
@ewwhite lol, it must be fun living in a nice big city. There are schools we support that can't get anything better then a T1
 
7:09 PM
Yeah, drop the T1, get DSL or wireless...
 
@HopelessN00b YMMV, of course. If I could get fiber, I would.
 
Good to know. Everyone around me's out of their freaking minds, not the other possibility.

@ewwhite I like how you think. Would you be willing to come out and tell some managers and execs that, employing the word "fuckwit" as often as possible?
 
The cheapest I could get fiber was $2k/mo for 10Mb sym.
 
@HopelessN00b I have a customer that needed to get internet access to a produce farm in Mexico...
 
@HopelessN00b Well, if they really WANT to pay for leased lines and the Five-Nines uptime...
 
7:11 PM
and this is the best we could come up with. Link balancer, USB 3G/4G, Mexican DSL, and Cisco ASA + VPN
 
I generally assume my company/client wants the best bang for their buck
 
rough, but it works... and is transparent to the users.
Hmm, I just learned dmesg -c
 
I get ~ 25Mbits/sec down on my 4G connection. If it weren't for the fact that I'd hit the data cap in about two hours, I'd use it all the time.
 
You learn something new every day as a sysadmin...
 
@MichaelHampton Right, I'd love fiber, but if they're unwilling to spring for even road runner business class, well, yeah.
 
7:15 PM
@HopelessN00b I suspect you aren't presenting the option properly. Remember that you should emphasize the cost savings and the bandwidth increase. :)
 
@ewwhite wot's -c? my dmesg doesn't respect such a flag
 
@voretaq7 clear
 
   -c, --read-clear
          Clear the ring buffer contents after printing.
 
@Zoredache why would you want to clear the message buffer?
 
@voretaq7 Clears the ring buffer. After all of the crap from the link up/down messages...
 
7:16 PM
@ewwhite they'll go away on their own eventually :)
 
@voretaq7 so you don't have a bunch of crap to wade through the next time you run dmesg -c. Besides it's all logged to a file anyway.
 
all of...
 
@voretaq7 because it is probably getting logged to a file already, why not clear the in-memory bit so you can read it?
 
bnx2 0000:03:00.0: eth0: <--- start MCP states dump --->
bnx2 0000:03:00.0: eth0: DEBUG: MCP_STATE_P0[0003610e] MCP_STATE_P1[0003600e]
bnx2 0000:03:00.0: eth0: DEBUG: MCP mode[0000b880] state[80008000] evt_mask[00000500]
bnx2 0000:03:00.0: eth0: DEBUG: pc[0800b578] pc[08008f40] instr[03623824]
bnx2 0000:03:00.0: eth0: DEBUG: shmem states:
bnx2 0000:03:00.0: eth0: DEBUG: drv_mb[0103000c] fw_mb[0000000c] link_status[0000006f] drv_pulse_mb[00002e42]
bnx2 0000:03:00.0: eth0: DEBUG: dev_info_signature[44564903] reset_type[01005254] condition[0003610e]
 
@MichaelHampton Possibly, but then neither is anyone else. I'm far from the first or the only one singing the same "sanity" tune.
 
7:17 PM
and those don't look like friendly messages.
 
@ewwhite @Zoredache :shrug: I just look at the leading dates :)
 
@HopelessN00b If you have executives who don't care about cost savings, run like hell; something's terribly wrong with your company.
 
@HopelessN00b present them a table with the features of each solution, and a price at the bottom
As long as both solutions are acceptable it doesn't much matter which they pick :)
 
@MichaelHampton Well, they make truckloads of money and pay far too well to do that. And, oddly, are spending millions on an infrastructure hardware refresh. All the new toys and systems go on my resume and will make me able to name ridiculous fess in a few years' time.

@voretaq7 Not a bad idea. Predicated on the shaky assumption that logic matters, but it's worth a shot.
I just... it's like... what hell? How many hundred thousand dollars did you spend on the SANs I unboxed yesterday, and today you're freaking out over $200 a month? Um, ouch, brain hurt bad now.
 
@HopelessN00b what type of job is this?
 
7:24 PM
Sr. sysadmin/architect/engineer
 
@HopelessN00b I still think you're approaching this the wrong way. Please go here and read everything on this page for some alternative ideas: bofh.ntk.net/BOFH/index.php
 
@HopelessN00b But you're a hopeless noob!
 
We bought a home! (To be built.) -- 360.io/Burf7Y
 
@HopelessN00b If you present the solutions and a recommendation to management and they decide to do something that is technically inadvisable make them sign a memo that says they understand this is not the solution you've recommended and that it has these problems [insert list], but they are proceeding with it anyway for these reasons [insert list].
 
@MichaelHampton right, exactly what I need. *More* encouragement to be a BOFH. To the execs.

@ewwhite You didn't ask what position I was qualified for, just which one I have. :D
 
7:30 PM
If they refuse to sign such a memo produce one yourself (minus the reasons), give it to them, and keep a copy. You've done your job by warning them.
 
@voretaq7 @hopeless because, you don't really get to say, "I told you so"....
 
@HopelessN00b Some execs count, others don't :)
 
@ewwhite no, but when they try to fire you for incompetence you produce the memo
(they may still sack you, but if they try to trash-talk you they come out looking like shit)
 
and they fire you anyway :)
 
@voretaq7 Yeah, that's my standard... CYA... M.O., but honestly, it never does any good. If you ever need it take cover behind, they blame you somehow anyway. Any time I've gotten to thinking about having to pull one of those, it's time to find a new job anyway.
 
7:32 PM
@HopelessN00b Yep, you've got the whole thing backwards. YOU are supposed to make THEIR lives a living hell.
 
I think it's important to make them know that you're hired to be the expert, or at least to distill expert information... They have no business second-guessing your work.
Or maybe they do..
 
@MichaelHampton Can't do that job. The current infrastructure does that aready.
 
@HopelessN00b @ewwhite It's not about the current job, it's about the next one when they say "Yeah, I heard about you - you're the dude who cost Shiny Widgets PTY LTD their entire production environment because you told them to use a dial-up modem" and you can turn around and show that you actually gave them sound advice which was ignored.
 
@voretaq7 Good point. Though I've never had the unpleasant experience of needing it for that either.
 
@voretaq7 I believe that's illegal in the US. (Though it probably happens anyway...)
 
7:34 PM
No, not illegal.
Opens the other company up to liability, so they generally don't say **** to a prospective employer of yours. Just a "not worth the risk of a defamation lawsuit" situation.
Or of they say good things about you, and you suck, the new employer could theoretically win a lawsuit on some grounds or another, so... just the lawsuit-happy US culture.
 
@MichaelHampton It's perfectly legal for two managers to go out to drinks and one Bob to say to the other Bob "Yeah, we fired our network guy because he's an idiot - you wouldn't BELIEVE the shit environment he designed for us..."
 
@voretaq7 Oh sure, like I said it probably happens. Just not through "official" channels.
 
For that matter your handing over the memo may be illegal (depends on the terms of your NDA)
 
I've seen it happen through official channels too, actually, if the guy's really that good or bad.
 
So apparently Intel's CEO has taken an interest in my "hot" issue.
 
7:43 PM
which issue?
 
Ouch, bad news. Never notice the CEO's wife or daughter is hot, because it becomes an issue real quick.
 
Wife, no. Daughter, only if you don't get caught.
 
nas-tay
 
2 days ago, by MikeyB
user image
 
Oh, Intel server... my server fire was my fault.
that sounds like a defect.
 
7:57 PM
@AaronCopley Looks like a nice area. If you've got a semi-nice camera, try using Hugin to stitch the photo, works freaking awesome for me.
 
Lol, that's hilarious.
 
@MikeyB "Issue Severity: Hot"
did they get back to you yet BTW?
 
@ewwhite I'm wondering if it has anything to do with the fact that the 1200W PSU's FRU information is set to warn at 1533W and trip out at 1750W. I measured it running full load of 1350W on that single PSU.
 
Chassis fires impact shipment... who knew?

Learn something new everyday. Have you suggested finding less whiny customers?
 
@voretaq7 Oh yeah, shipped it to them yesterday, priority overnight to .wa.us. They wanted to get their hands on it rather quickly.
 
7:59 PM
@MikeyB Usually that stuff has a 1:2 safety margin....
 
@MikeyB uh huh.... yeah... and? . . . IT WHAT?!?!?!
@ChrisS that was back when actual engineers were writing the spec sheets
 
@voretaq7 Oh, right Cloud Readyâ„¢ version is Marketing approved.
 
@ChrisS approved?
that would imply it gets "released"
this is THE CLOUD man! Software just IS!
No "releases" no "version numbers" man - what are you, some kind of a waterfall-model FASCIST man?!?
 
@voretaq7 That's pretty much it. "Yes, it literally caught on fire."
 
Well if someone would put my in charge I might be fascist... But at is I'm more of a peon.
 
8:06 PM
Sigh… I asked our padawan tech guy to compile a list for me of something that's easily tabular data.
Yes, he put it into a spreadsheet (he passes that test).
No, he didn't email the spreadsheet. He emailed a PDF of the spreadsheet.
2
 
@MikeyB Death is too good for them.
 
@MikeyB Was the generated PDF 3x bigger than the original XLSX file?
 
@ChrisS 15.5KB of PDF. What would a smallish .xlsx run?
 
That, more or less. A blank xlsx is 8.15kB
 
I'm trying to decide which I hate more: debian or OSX.
Every morning I'm like, "It's definitely X," but subsequently every day I find something new to hate about Y
So it's really a toss-up where I hate Debian and OSX each more and more, in equal proportion.
 
8:14 PM
@JoelESalas Debian.
 
@MikeyB Did he paste a screenshot of the pdf into a word doc and send that?
 
OS X never broke OpenSSL.
@MikeyB Was the PDF at least a proper PDF table that you could cut and paste?
 
@JoelESalas OS X is nice to look at. Hate Debian more.
 
@MikeyB Have you considered marketing those chassis fires as a security feature?

You know, to self-destruct the data in case of theft?
 
@MichaelHampton I have to turn all that shit off to get any work done. It's like a less-useful, not-standard XFCE install
 
8:16 PM
@JoelESalas OS X only frustrated me because ⌘-Tab switched between applications and not windows. As soon as I discovered tmux, this was much less of a problem.
Of course, I'm back on Fedora...
 
@HopelessN00b Sorry, my co-worker already won this contest. He sent the pics to me with the subject "Intel HCF error".
 
@MichaelHampton Fedora master race
 
And using KDE too. None of this GNOME 3 crap.
 
@MikeyB 's not an error. The instruction seems to have worked exactly as designed.
 
@voretaq7 I suspect a careful reading of the logs will indicate that it caught fire first and then halted. In other words, it performed the HCF instruction in the wrong order.
 
8:23 PM
@MichaelHampton Nothing in the instruction set documentation says that HCF is atomic
(or deterministic)
 
@voretaq7 I'm having trouble finding docs, but it seems the original implementation halted first, and then caught fire.
 
@MichaelHampton Well, considering that the servers halted (and powered off) Fri. night and caught fire Tue. afternoon (with the servers themselves off), I'd say the order was correct.
 
@MikeyB Ah, I missed that detail. In that case, it seems to be working fine.
 
@MichaelHampton There is ample documentation of computers continuing to run while on fire - though whether this is because of a flawed implementation of the instruction or not is difficult to determine
 
I am reminded of Jun 25 12:20:47 pc1h kernel: lp0 on fire, which as far as I know, never worked right; nobody ever reported an actual printer on fire...
 
8:32 PM
@ChrisS I had a Canon 5D Mark II / 17-40L with me. But no CompactFlash cards... LOL
FFFFFFffff...
 
@MichaelHampton raises hand
we didn't get a console message about it though
 
Writing an interview apology letter...
 
(shitty old HP3 -- paper jammed up in the fuser and... well...)
 
Afraid to click SEND
 
@ewwhite "Dear Sir or Madam (frankly I'm still not sure -- it's the hair): Sorry for embarrassing you like that during my interview, but frankly you're storing your company's production data on a RAID 0 of 10 year old PATA drives, in a tent, in the parking lot. You're absolutely batshit bonkers. . . . ."<- ?
 
8:39 PM
@ewwhite a ..what?
 
@ewwhite Wow, did they need you there yesterday?
 
@ewwhite pfft, have 'em put you up in a hotel man!
 
King of redaction, thy subjects await your appearance.
 
Well, the other offers gave me a work-from-Chicago option.
 
@voretaq7 Everything sucks. (Axiom) Everything is. (given) Pie! (because)
 
8:42 PM
@WesleyDavid I refute your given - Show how you derived it from existing axioms/theorems!
 
@voretaq7 I would have liked that... but I had things to sell in Chicago. It's not a small thing.
 
@ewwhite Storage company and Craigslist :)
 
@voretaq7 Two condos.
 
@ewwhite Storage company, Craigslist, and a realtor :)
 
@voretaq7 pokes you in the eye That hurt? Proven!
 
8:44 PM
@WesleyDavid stuffs you in a box and leaves you in the tape safe What kitteh?
 
getting ready to click send...
 
And now I want pie.
@ewwhite Do it. Do it! DO EEEEETTTT!!
Want me to proof it?
 
@voretaq7 Finding myself poor and penniless on the tough streets of NY...
 
@ewwhite That NY Times reporter made a good living panhandling for a week... and it's tax free!
(or at least damn hard to track unless you get greedy and become one of those high-tech bums with an iPhone credit card reader...)
 
Hmmm, good hustle.
It seems weird to apologize, though. But I feel bad.
The job was reposted on Careers.SE.
 
8:52 PM
@ewwhite Alternate answer: "...Is better than those idiot college students who go all the way to Europe to find themselves and wind up penniless on the streets of Prague or something!"
 
Is my desire to be a wealthy sysadmin completely crazy?
2
 
@ewwhite OH MY YES!
 
bubble burst
 
Wealthy? I think the best you can hope for is to always have the latest kit to play with.
 
Don't talk to me about scary moves to NYC... O_o
 
8:57 PM
@MichaelHampton Heh, I've done okay. But I'm becoming more demanding in my old age.
 
@voretaq7 Was this the story that was published a few years ago?
 
@BartSilverstrim That is true... I couldn't pull it off when I had another company that was more flexible about it.
@bart how is it going? Where did you find an apartment?
 
59th street.
 
@BartSilverstrim it was a while back yeah
 
Location was actually pretty damn good. The apartment is...not quite as good, but the location I like.
 
8:59 PM
heh... dare I ask about $$?
 
Might be better off dressing up as Elmo on the weekend and harassing tourists in Times Square.
 
@ewwhite I thought the only people that got wealthy these days, controlled money in some semi-direct way or real estate.
 
@ewwhite What do you mean, the rent or the pay or the ...?
 
@BartSilverstrim CTW would sue your ass into the ground :)
 

« first day (698 days earlier)      last day (4298 days later) »