The Comms Room - 2013-05-02 (page 4 of 5)

« first day (964 days earlier) ← previous day next day → last day (4025 days later) »

7:00 PM

@Travis Yes, i'm still looking at the thread you forwarded: some of the links in there are broke. And I'm just going to assume UAC disable at this point

@Clickinaway It's a method, although not the best method. The link just had a lot of different things to try. I wasn't sure all you had tried so I wanted to try to narrow it down.

Woo! It's Noon and it's already 70F!

@Travis hmm...there's a little nugget in this though which I hadn't thought about doing and i'm not sure the case would work in PY (using 'run as' at the command prompt) itberius.blogspot.com/2011/11/from-httpriosec.html

@freiheit Not surprised. They also use Confluence.

@Clickinaway Yep. That was where I was going when I asked about run as admin.

7:08 PM

@Travis I had just been thinking about it in the context of the GUI/Desktop

@Travis awwww yeeeah

back to PY now

Michael Hampton

What can I say, Server Fault is just a bunch of surly curmudgeons. — Michael Hampton 8 secs ago

bummer

no way to put PWD in the 'run as' CMD?

@MichaelHampton That makes me wonder who he's pissed off in his other questions.

Michael Hampton

@ScottPack I dunno. I was about to go looking.

That was, after all, nowhere near my best answer.

@MichaelHampton Not that your answer is bad, but it's not exactly a shining example for all to aspire to either.

Michael Hampton

7:19 PM

@ScottPack Exactly. I have yet to write any answers like that.

@Clickinaway Yep :) How about this. Use PSEXEC with credentials to run the original PSEXEC that executes the bat file?

@MichaelHampton @ScottPack the answers to his other questions all appear to be quite polite

maybe he just needs a kitten to hug

Michael Hampton

Yeah, I didn't see anything out of the ordinary in any of the ones I looked at.

@voretaq7 @WesleyDavid alright get your mangy ass in here.

@voretaq7 Hello (=・ｪ・)

7:26 PM

@Clickinaway OR, use PowerShell and run it with credentials to execute PSEXEC. Here's a sample: blogs.technet.com/b/benshy/archive/2012/06/04/…

FINGARS!!

@JoelESalas Is it can be hugs tiem now?

@WesleyDavid :3

HUGZ TIEM!

@gWaldo {{{ @gWaldo }}}

user image

^ This is why I love fedora

7:28 PM

cough GENTOO

GET OUT

@JoelESalas That's pretty funny

@JoelESalas Should I make a cflags reference or do you want to?

Michael Hampton

@Olipro emerge me a sandwich just doesn't work as well.

sudo !!

7:29 PM

silly, it's emerge sandwich-maker

Oh noes! The EMC SAN failed!

Michael Hampton

@ewwhite Is this the one your predecessor misconfigured?

@ewwhite But, EMC is on CSI it can't fail!

@MichaelHampton Current coworker.

Michael Hampton

@ewwhite Who admits nothing, of course.

7:31 PM

Multipath issues.

multipathd: sdc: readsector0 checker reports path is down
kernel:  rport-5:0-0: blocked FC remote port time out: saving binding

user image

Your Personal Rate of Return
This Period 12.0%

Yeah what's up, whatchu know retirement account

@JoelESalas whaddya want? I'm dealing with fucking Linux today...

Once again, Linux inspires nothing but rage and hatred. If this collection of half-assed hacks is supposed to be "The Future" we're FUCKED.

Tweeted by voretaq7 on May 2, 2013 at 7:13 PM

@JoelESalas Stupid VTEC. It's really quite obnoxious on Honda's motorcycles.

@JoelESalas how long is a period?

@voretaq7 YTD

@Adrian Drives me nuts on my Accord too, which is why I'm ditching it

7:39 PM

@JoelESalas Hmm. VTEC on the regular Accord doesn't annoy me anywhere near as much as the crappy CVT with battery boost that Honda has the temerity to call a Hybrid

I think I got 31mpg out of the "Hybrid" Accord that my Lady borrowed a few weeks back.

@JoelESalas hmm, can't figure out how to extract that value from Fidelity's shitty web interface

@voretaq7 workplaceservices200.fidelity.com/mybenefits/navstation/…

@Travis PSEXEC within PSEXEC? wow..is that the blue or is that the red pill? as for PS option just read that but not seeing a place to store credentials? remember i'm stuck with this particular local account (even with admin privileges) as the initiator of all processes.

@JoelESalas my account is an IRA - that page won't work for me

:(

@voretaq7 Why'd you do an IRA? I'm still learning the ins and outs of this

7:42 PM

I know I lost $40 today, and have gained substantially more than that since inception

Michael Hampton

@voretaq7 Got a problem with Linux? Install FreeBSD. Then you'll have two problems.

2

@JoelESalas because I've had one since I started working, and the 401(k) from $job[-1] rolled into one when I left.

@voretaq7 Roth or Traditional? Also, rationale?

I'll be able to do a goddamn fucking OS upgrade without having to visit 100 fucking machines in the fucking field.
FUCK FUCKING LINUX!

@JoelESalas Traditional because I can use the tax deduction right now.

Michael Hampton

@voretaq7 Why would you have to visit a machine to upgrade it?

7:43 PM

I may open a Roth at some point to throw extra money into

@MichaelHampton because they're laptops, at client sites, and the client can't have root.

Michael Hampton

@voretaq7 It doesn't sound like Linux is the problem here.

and Linux does ridiculous things like change the parameters to grub-install with NO BACKWARD COMPATIBILITY so the upgrade scripts that worked last week don't work anymore, which would leave every field unit unbootable.

Michael Hampton

Eh? grub-install has more than one parameter?

@MichaelHampton it has one mandatory parameter (the disk to install to)

the format for that parameter used to be '(hd#)'

Michael Hampton

@voretaq7 Hmm, that was ages ago. And pretty stupid.

7:46 PM

that format is no longer valid - it MUST be /dev/[whatever]

@voretaq7 I'm in the exact same scenario with the 401(k) I have right now

@MichaelHampton stupid though it may have bin, it was the ONLY way that worked reliably up until the latest release of Ubuntu, so it's the one my upgrade process relies on.

Michael Hampton

@voretaq7 Ohhhhhh, Ubuntu. I think I see the problem now.

Now that the wonderful Linux Developers have broken that, my entire upgrade process is in the shitter because it leaves the machines unbootable (can't install the new boot loader, so it drops you to a grub rescue prompt which is the single most useless thing on earth)

@MichaelHampton Ubuntu is just providing the (broken, zero-backwards-compatibility, no-notice-of-deprecation) Grub package the Grub people produced

so for once I don't want to cockpunch Shuttleworth (well no, I still do, but I also want to cockpunch the entire GRUB team...)

Michael Hampton

@voretaq7 GRUB 2? That's been pretty widely known to be coming in the Linux universe for years. It got to the point where I wondered if it would ever arrive...

7:49 PM

@Clickinaway lol. What I was figuring is with runas you can't specify the password. With psexec you can and since your account has local admin rights you run psexec with local admin right context but also specify account credentials to run the 2nd psexec on. So it would be something like 'psexec -u user -p password "psexec \\computer ...."

@MichaelHampton yes, grub 2. Where functionality just disappears instead of being deprecated and issuing warnings.

Michael Hampton

@voretaq7 It should be done the Microsoft way, then?

@votetaq7 Choose a distro that mostly maintains the state of the packages original intent.

@MichaelHampton dude - I would seriously rather be fucking shipping Windows right now

that's how much of a pain in the ass this is

@MIfe this isn't a distro thing. It's a "The grub people removed functionality without deprecating it and issuing warnings first" thing

@votetaq7 What makes you believe the burden is on them? Surely your distro should be keeping track of these changes.

7:52 PM

@MIfe because they're the vendor responsible for the grub-install shell script

and I as a user should not be put in a position where core functionality disappears without a period of "X is deprecated and will go away soon" warnings.

Michael Hampton

@voretaq7 But... but... they changed the major version number!

OK, so why did your distro think it was a good idea to change the package in the first place?

@MIfe Because grub 1.x is no longer supported by the grub people

@MichaelHampton Indeed - which is the right thing to do. But can we not also agree that you tell your users about the change and provide for a transition period?

Right, so then the distro if necessary maintains it as much as possible, or moves people having given prior warning.

Michael Hampton

@voretaq7 Four years wasn't enough?

7:55 PM

Out of curiousity, was this an in-place update (I.E ubuntu12 to ubuntu12 + updates) and not like ubuntu11 to ubuntu12?

@MichaelHampton when was there ever a warning that (hd#) was deprecated in the last 4 years?

Michael Hampton

@voretaq7 Good question.

@MIfe 12.04 to 13.04 -- to Ubuntu's credit their upgrade script handles it. Mine chokes on deploying the changeset (because I relied on what the grub people told me was the preferred syntax in grub 1.x, and that preferred syntax disappeared entirely in 2.x)

@voretaq7 I could go for a comforting hug.

@MichaelHampton And I would be slightly less pissed about this if /dev/whatever and (hd#) were functionally equivalent, but they're not

the "(hd#)" syntax did magic to figure out what the "first BIOS drive" was, and would work in mixed hdX/sdX environments.

7:59 PM

@Travis then supply credentials again? my brain keeps going through thought loops on that one...will have to remote back into work tonight i guess

shit like this is why I keep saying Linux is not production-quality :-/

2

Linux is a moving target. thus if you want a stable platform you choose a distro that maintains some remnant of stability.

@MIfe EVERY OS is a moving target

@voretaq7 I won't pretend to know enough to discuss this. People do run businesses on Linux though.

People also run businesses on NT, apparently. (@Cole)

@votetaq7 I would agree there.

8:01 PM

but no other OS I've ever used is so godawful at maintaining continuity that the standard way of dealing with upgrades is to reinstall the entire system (I'm looking squarely at RedHat when I say this).

Define 'upgrade' then.

@voretaq7 Right, but it's a completely different philosophy. In Linux-land, trashing a machine isn't a failure of the system per se.

@voretaq7 :(

One doesn't reinstall the system to go from say 6.3 to 6.4.

@MIfe We're talking 5 to 6 for example

8:02 PM

@JoelESalas yes it bloody well is! -- If an upgrade requires you to physically visit a machine and lay hands on it you've failed.

you just don't do that. Well I did once.

to continue on the previous idea, most "cloud"-type architectures are predicated on the idea that your servers aren't special snowflakes and your application should be distributed enough to trash any given server at any time without too much trouble

@JoelESalas 5 to 6 is a full reinstall... and it'll kill your performance.

@MIfe In almost 20 years of working with FreeBSD I have had to physically touch a machine to do an upgrade exactly once. And that situation was treated like the end of the world by the entire FreeBSD community.

@JoelESalas yes, but that's bullshit in real life. I NEED to have snowflake systems.

@JoelESalas OK, what if instead of servers I have 1000 workstations in my Fortune 500 company?

Michael Hampton

8:03 PM

@voretaq7 Just how ancient are these systems?!?

@ewwhite I keep trying to kill all the snowflakes, but they just keep coming

@ewwhite Right, but that's a business need as a function of the business model and industry in general

@MichaelHampton Some of them do have /dev/hd# devices (early 2000s) -- it's not a major problem for me to account for those though

@freiheit All of my produce systems are one-off ERP systems. The world doesn't run on clusters of web servers...

@voretaq7 Seriously though, I've only had the kind of upgrade problems that you so frequently and colorfully describe on Ubuntu.

8:04 PM

@JoelESalas @voretaq7 I think thats a reasonable expectation -- you cant have your cake and it eat. If you expect the platform to remain stable then you cant expect it to work as an 'upgrade' because that stability that you just hinged your entire environment on is gone.

@ewwhite yet

@MIfe RHEL used to support version upgrades.

@ewwhite Some of us work at hip young startups where everyone wears a hoodie

@JoelESalas and you end up running websites for. ever.!!

@MIfe I don't think it's reasonable to make compatibility-breaking changes without a deprecation period. And there is no way on God's green Earth that you will convince me "reinstall from a CD" is an acceptable upgrade process for Unix systems because there's 40 years of history where in-place upgrades are the norm and having to lay hands on a box is considered a shocking exception

8:06 PM

@ewwhite I actually want my systems to remain on a stable API in this context. I would not consider performing an 'upgrade' like that safe.. even if we assume that the process is 100% compatible between package sets that are being upgraded from RHEL to new RHEL -- I'm not able to say my configuration for my services will work, that my applications I developed will operate correctly or perform in a way that I expect.

user61389

I have a very low-level question about the TCP protocol. My PC talks to a chip server, which replies, but the PC (client) keeps asking for new TCP connections, so I wonder if there's something wrong with the chip's reply. Would such a question be on topic here?

Michael Hampton

@MIfe Funny, that's what Red Hat said.

@MIfe OK but conversely if I test my application and confirm it to be fully functional, shouldn't I be able to do such an upgrade?

@MIfe well, that's where a form of configuration management helps on your snowflake systems... you should be able to rebuild it onto a new platform.

test in parallel, migrate data.. be done.

@ewwhite I have no shame about coming to you in rags and failure, begging to puppetize your produce servers

Michael Hampton

8:08 PM

@CamilStaps Unless you actually know what's going on, you'll be told to "Fire up Wireshark and look at what's going over the network."

@JoelESalas what's to puppetize? I run them for 3-5 years and upgrade.

@voretaq7 and how would one erm, test it? By putting it on a new host with RHEL6!

user61389

@MichaelHampton that's what I'm doing, but I'm not familiar with the protocol. I can include the packets that are sent and received in the question.

@ewwhite I guess I need to revise my plan then.

Michael Hampton

@CamilStaps That would definitely be helpful.

8:09 PM

@ewwhite that sounds so reasonable and sane (...but my configuration management system's migration process was just broken by an upgrade! Oh sweet irony!)

user61389

@MichaelHampton and it is on topic?

@MIfe by performing the upgrade in a test environment

Michael Hampton

@CamilStaps That depends. Are you getting paid for this?

user61389

@MichaelHampton haha, no, it's a hobby project

OFF-TOPIC'D!

8:10 PM

@CamilStaps Never ever admit that.

Michael Hampton

@CamilStaps You'll probably be run off, if anybody finds out it's not a professional project. So don't mention it. Or just throw up your network captures here in chat and maybe someone will sort it out.

user61389

@MichaelHampton hmm okay. Would there be another stack site for this then?

Michael Hampton

@CamilStaps I dunno. I'd have to see the actual question.

@voretaq7 Personally I consider in place upgrades of that nature when coupled with proper testing to be -- probably -- about equal amounts of effort.

@CamilStaps Technically SuperUser, but they might not have the TCP/IP depth to answer it

8:11 PM

Clearly you never tested in your circumstance else your gripe would have been known ;).

user61389

@MichaelHampton then I'll just ask it here and if not perhaps it can be migrated?

@voretaq7 So you're admitting that you didn't do a QA canary test with the new ubuntu and your config mgmt.?

@MIfe my circumstance is 13.04 was just released last month and I'd like to upgrade to it BUT I BLOODY WELL CAN'T BECAUSE THE BROKE SHIT.

@Adrian this is the canary test.

I suggest you do the tests you mentioned first then :).

^ And that's the fucking result ^ :-)

Michael Hampton

8:12 PM

@CamilStaps Yeah, just ask the question, and make it Really Good.

@voretaq7 yeah, I wouldn't be doing canary testing on equipment that I don't have physical access to. Even if you have to simulate the connectivity coming in from outside.

@Adrian I do have physical access to my test systems. That's how I determined it was the grub install script causing the problems

I'm enjoying the buttmad

@voretaq7 Yes, and this happens all the time with Ubuntu. I'm really quite surprised that you didn't already know about that.

@JoelESalas oh this is assrage dude

@Adrian it's not just Ubuntu (they're just my personal hell)

8:16 PM

I dont keep with ubuntu (more of a redhat type of guy personally) but dont they offer like LTS releases?

@voretaq7 I've done similar upgrades with Debian where it worked just fine.

@MIfe yeah - 12.04 is a LTS release. I'm trialing this upgrade because of problems with the LTS release

Michael Hampton

@MIfe Yes, but that's no guarantee of anything. Ubuntu's upgrades are broken more often than not.

@MIfe Yes, but the cycle is still only 3 years.

@voretaq7 lol. Not that long term I guess :)

8:17 PM

@Adrian @MichaelHampton This would work fine if I could give my end users root access and just use the do-release-upgrade script

@MIfe Unfortunately, Ubuntu tends to do really unpleasant things with their LTS releases IN ANTICIPATION of the LTS aspect instead focusing on long-term supportability

When they released 10.04, they totally changed around the GUI and integrated a bunch social media stuff at the last minute. And it was a cluster-fsck.

@Adrian it's not just that - the LTS releases are really light on the "S" part.

12.04 was so full of memory-hogging crap that it effectively wasn't usable where I was working at $job[-2]

yes they keep shipping patches, but they don't incorporate major fixes

@voretaq7 I'm pretty sure that describes every single aspect of Ubuntu.

8:19 PM

maybe you can use sudo voteraq or is it like company policy?

so you wind up with 10.04 that by 3/2011 didn't support any kind of reasonably new hardware

Vote Raq!

@MIfe it's "I can't let Ubuntu just install whatever it wants" policy

@voretaq7 Yeah, you should try using their crap in large multi-user use cases. It's a fiasco. If it's not a single-user desktop, they care even less than the already ridiulously tiny bit they normally might care.

@MIfe The Ubuntu upgrade process throws up a bunch of prompts that require user input - I can't count on them following instructions and answering every prompt correctly. That's why I have to upgrade the template machine and then push the new template out.

8:21 PM

Oh, is there no silent 'here are the answers' thing?

When DBUS was crashing in multi-user setups in 2010, Canonical actually told a guy that had a support plan to just reboot his servers weekly instead of actually diagnosing what was going on with the 2-3x/month DBUS lockups he was getting.

@MIfe not if I use Ubuntu's do-release-upgrade process :-/

Michael Hampton

@MIfe Yes, you can use the debian noninteractive trick, or preseed the answers.

@voretaq7 Oh, that shit.

@Adrian that's..... pretty much the standard for Linux in general. Not. Production. Ready.

@MichaelHampton yeah - if I could just ship all my users CDs or USB sticks this wouldn't be so much of a problem

@voretaq7 Actually when you have redhat support they are pretty good at actually giving a shit i've found.

8:23 PM

@MIfe unless you want to do something like upgrade 5->6 or 6->7 :)

I refer you to my previous responses ;)

@voretaq7 Oh, is that why Google runs on Windows? Oh wait...

and I've had RHEL support cases where the answer was "Yeah, that's fixed in the next major version" when we're talking about a 2-line patch to a file...

Michael Hampton

20

A: Why is it so difficult to upgrade between major versions of Red Hat and CentOS?

To start, I should note that there are two ways to do the in-place upgrade: Drop in the installation DVD (or use the DVD image via iLO/iDRAC), boot from it and choose Upgrade, e.g. linux upgradeany. Update the redhat-release RPM manually, run yum distro-sync (this is oversimplified a bit) and r...

@Adrian (a) there's more Windows at Google than people think. (b) Google has the advantage of being able to PXE boot everything and reinstall with kickstart/preseed/etc.

8:25 PM

@voretaq7 yeah that is a issue. I'd push there on that. If they are saying they support long term they should backport bugfixes.

Michael Hampton

So, is there someone you can pay for the privilege of yelling at them if FreeBSD breaks?

If you just report a bugfix and dont get support they have every right to say that, but if you're paying for a platform that has bugs that have known patches that can be backported they should be.

Michael Hampton

@MIfe But... having the same bug for 10 years is enterprisey!

If your asking for new features yeah thats not what they offered with the platform anyway.

@voretaq7 Yes, I know. I have a number of friends that did or do work there. But they're also running about a gazillion Linux VMs under KVM there.

8:26 PM

@MichaelHampton iXSystems is the usual answer

@Adrian ...with the ability to PXE boot the hosts and reinstall them from a kickstart configuration.

DevOps == Hats?

user61389

Okay then, let's see what happens :)

user61389

0

Q: Is this TCP connection correct?

I'm building a server with the ENC28J60 chip and a PIC18F4620. The chip is connected to my PC via ethernet. Currently, I'm trying to set up the TCP connection over which I'll build an HTTP connection later on. I have never worked with TCP before. After the needed ARP requests and replies are sen...

tcp ethernet socket client-server

@voretaq7 Yes, because they're using a production-ready distro, not the POS Ubuntu stuff.

@Adrian I could do the same thing they're doing with any OS

it's really easy when you can put your hands on every machine (or the equivalent with PXE boot and virtual CDs over the network)

it's a lot harder when you have to let the shit out of your network and make it work without remote hands.

8:29 PM

@freiheit @voretaq static.pistoncloud.com/images/team/lloyd-dewolf.jpg fedorasofokc.jpg

@voretaq7 Of course it is, but you've also picked one of the hardest distros to support.

@Adrian what would you suggest as an alternative? Bearing in mind that every distro that went from grub1 to grub2 is going to have the issue I'm dealing with now (which is why I'm cursing all of Linux and not just Ubuntu this week)

@voretaq7 OpenIndiana. sees himself out

@JoelESalas A trilby is more devops than a regular Fedora

Michael Hampton

@CamilStaps Are you writing a whole TCP stack yourself?!

Dennis Kaarsemaker

8:31 PM

@voretaq7 backwards compatibility is for pussies!

user61389

@MichaelHampton very low-level!

@DennisKaarsemaker You know what else is for pussies? Meow mix.

Dennis Kaarsemaker

@WesleyDavid pussy needs to buy a bike!

3

Michael Hampton

@CamilStaps I mean on the "PC".

@CamilStaps I think the ack your getting back looks pretty anomalous. I tried a simple http connection to google and the ack values nowhere near that number. Althoguh i'm not sure if what tshark is giving you back is the absolute ack value or a relative one.

8:32 PM

@WesleyDavid ...that's not a bad suggestion.

@voretaq7 To be honest, I don't much care for Linux on the Laptop. I'd have left them on Windows. Too many issues, too little support from vendors, for too little return.

@DennisKaarsemaker You're so mean!!

=P

@Adrian It's the year of the Linux desktop OS!!

And for this morning's dose of heresy, I miss my MacBook Air from $job[-1]. =(

GRUB update kills grandma's laptop Well shit.

@Adrian yeah, well that choice was made by the developers. At this point I'd rather be shipping Windows myself.

and we all know how I feel about Windows.

8:35 PM

YFa25uB5XP3r2isasawm1hfcY9mD9qFrw4j!

Good VPN key?

@ewwhite Awfully short.

@ewwhite absolutely. Use it on all your highest-security networks.

:-)

Michael Hampton

@ewwhite It was a good VPN key until 30 seconds ago.

Edits...

user61389

@MichaelHampton oh, no, I'm doing nothing with the PC, that's just entering the IP in the browser and it works (well, that's what should happen)

8:35 PM

@ewwhite also how the fuck did you get my root password?!

user61389

@MIfe what do you mean? I'm not familiar with TCP, nor with wireshark

@voretaq7 Under your keyboard.

@voretaq7 My password generator... Horse Staple battery...

Dennis Kaarsemaker

@voretaq7 but your rootpassword was hunter2!

@ewwhite That's some kinky stuff right there. Didn't a guy in Seattle die doing that?

8:37 PM

@CamilStaps en.wikipedia.org/wiki/…

@WesleyDavid Enumclaw. About 40 miles outside of the city.

wikipedia says that you should get back a value of the syn+1.

Dennis Kaarsemaker

@ewwhite I'd have expected Melons Fruity Peaches for you.

Michael Hampton

@WesleyDavid No, that was Horse Goatse Receiver.

Dennis Kaarsemaker

Not sure whether that's because of the produce jobs, or brazzers.

8:38 PM

@Adrian Drinking buddy, right? =P

@ewwhite

for i in {1..4}; do perl -e 'rand($.)<1  && ($line = $_) while <>; print $line' <(perl -nle 'print if m/^[a-z]{3,10}$/' /usr/share/dict/words); done

@WesleyDavid lies! (it's under my MONITOR!)

@WesleyDavid No, more likely my Ex's new husband's cousin.

Michael Hampton

@CamilStaps OK, so you're implementing TCP on the embedded device, then? I'm a little confused here.

@freiheit for theLoveofGod in {heaven};....

user61389

8:39 PM

@MIfe thanks. So the first steps are right with the flags, but the third isn't. However, there can't possibly be something wrong with the client, right? So there must be something wrong with the numbers A or B - am I correct?

user61389

@MichaelHampton yes indeed, let's see if I can clarify the question a bit

Michael Hampton

@CamilStaps Good idea.

@CamilStaps The first packet should have the ack value 1 from the chip.

Relatively speaking from what wireshark is producing.

not the actual value 1. But peers_syn+1.

@voretaq7 That's a really good place to keep your password!

user image

Michael Hampton

This is starting to look more and more like a Stack Overflow question.

user61389

8:42 PM

Ahh, I see. The client sends sequence number 0x5ef62198, the server replies with 0x5e002199. So the second byte is incorrect.

user61389

Now I'll only have to look at the code where I lose that one byte.

user61389

@MichaelHampton would SO deal with such not-directly-related-to-programming questions? I mean, it's about the protocol rather than the code.

@CamilStaps I think the intention here is you're implementing the protocol in code. So its a SO question.

Michael Hampton

@CamilStaps Generally, I think so. Though if you included the code, they'd probably also point out the bug for you :)

user61389

8:45 PM

@MichaelHampton it's a very very long code :) but I didn't know about that, what shall we do with this question? Now that we know the problem (for which I'd like to thank you a lot both!) it's very localized. Shouldn't I just remove it?

Michael Hampton

Naa, I wouldn't delete it. Though I suppose @MIfe could answer it and you could accept it.

Meh. I care about the challenge to answer, not the points.

user61389

@MichaelHampton yes, that seems fair to me

user61389

@MIfe then, as I don't think it would be useful to other users, should I just remove it?

Yep, fine by me.

8:48 PM

@CamilStaps are you basically reimplementing TCP?

user image

silly question...is there a setting on google that will actually show the full title and actually use some of the whitespace?

Michael Hampton

@voretaq7 Yeah, in his little embedded chip. That's why I voted to migrate to SO

user61389

@voretaq7 I didn't want to use Microchip's (my chip's manufacturer) TCP/IP stack functions, so I write them myself

@CamilStaps . . . because you like pain?

user61389

Anyone not OK with removing?

8:50 PM

(seriously, on my list of things to do reimplementing TCP is slightly below "deal with Linux upgrades")

Michael Hampton

@CamilStaps You probably should use them, unless you're just trying to teach yourself, or they're especially bad. In the latter case, get some other chip :)

user61389

@voretaq7 I try to use as little Microchip as possible. It's hardly documented. To be honest, I'm using Jal libraries (other language) as the base for this, so I don't have to do everything myself.

@MichaelHampton If the goal is learning write a TCP stack for Minix -- you'll cover the same ground, with less pain than embedded development :-)

Michael Hampton

@voretaq7 Minix? Does that even run on modern computers?

@CamilStaps It's hard to really screw up a TCP/IP stack (as a device manufacturer. It's really easy to have mistakes in one if you're writing it yourself.)

@MichaelHampton yeah - has SATA drivers now too.

Michael Hampton

8:52 PM

Wow. Next thing you know, GNU will actually release the Hurd.

@MichaelHampton . . . please, no. No more shitty software masquerading as a usable OS.

Michael Hampton

@voretaq7 But it's not easy to take down a multi billion dollar corporation!

user61389

@voretaq7 it probably works, but I have no idea how and it's hardly documented. Also I couldn't find example code for my chip family, etc. etc. this is a hobby project in the end anyway :)

@MichaelHampton Download a GNU/Hurd CD now!

@CamilStaps Round these ServerFault parts, we shoot hobbyists on sight. ;)

@freiheit no, we feed them to the monitor lizards :)

lizards gotta eat ya know!

user61389

8:55 PM

@freiheit I removed the question, you can't downvote me! :D

Michael Hampton

deleted by Camil Staps 2 mins ago

user61389

Anyway, I'm leaving. Thanks a lot for your help, anyone who helped! :)

Ugh figuring out my money is depressing

Do you guys always segment your networks?

I have a new 3-host ESXi setup, and am wondering what I should IP everything with.

« first day (964 days earlier) ← previous day next day → last day (4025 days later) »

Transcript for

May '132

The Comms Room

ServerFault's lobby

join this room
about this room

00:00

06:00

12:00

18:00

all times are UTC

site design / logo © 2024 Stack Exchange Inc; legal

mobile