Back to Home / #linode / 2006 / 03 / Prev Day | Next Day
#linode IRC Logs for 2006-03-27

---Logopened Mon Mar 27 00:00:31 2006
00:40|-|spr [] has quit [Quit: Spoon!]
01:21|-|avid [] has joined #linode
01:59<@caker>anyone on the xen waiting list and at HE want to move over now?
02:00<warewolf>just a bit? :)
02:00<@caker>been a LONG week, I'm ready to get this sucker going
02:00[~]warewolf is back to working on integrating a wiki into RML
02:01[~]warewolf really should be working on WORK stuff at work, and not RML
02:02[~]caker tries a kernel compile inside this new node
02:03<@caker>100%[====================================>] 40,845,005 8.86M/s ETA 00:00
02:04<@caker># time tar xfj linux-2.6.16.tar.bz2
02:04<@caker>real 0m33.707s
02:04<@caker>user 0m20.820s
02:04<@caker>sys 0m2.830s
02:04<fo0bar>caker: I just got 2.6.16 compiled on finnix
02:05<fo0bar>it seems OK so far
02:05<@caker>fo0bar: cool .. change your mind about including it in the next release?
02:05<fo0bar>I'm not sure
02:05<fo0bar>I'm planning for a release next week, so I've got a little time
02:06<fo0bar>I'm using the patchset from debian unstable, so I'll have to dig through to see what they've added/fixed
02:07<fo0bar>generally, I trust debian builds... the primary reason I announced 2.6.15 for finnix 87.0 was because it generally takes debian at least 2-3 weeks to come out with a new kernel build in unstable
02:07<@caker>any reason for using that patchset?
02:07<fo0bar>I fall under the camp that says "these distros don't modify the stock kernel just because they're bored"
02:08<fo0bar>oh, the kernel ppc -> powerpc transition seems to be complete
02:08<fo0bar>23:21 <%fo0bar> OK, in 2.6.14 you had ARCH=ppc and ARCH=ppc64, and life was good
02:08<fo0bar>23:21 <%fo0bar> in 2.6.15, for 32bit, you did ARCH=ppc, and 64bit was ARCH=powerpc
02:08<fo0bar>23:22 <%fo0bar> note, you could do a 32bit flag on ARCH=powerpc, but compilation would fail shortly after you began
02:09<fo0bar>23:22 <%fo0bar> now in 2.6.15, you can do 64bit ARCH=powerpc, 32bit ARCH=powerpc, and 32bit ARCH=ppc
02:09<fo0bar>23:23 <%fo0bar> err 2.6.16
02:09<fo0bar>23:23 <%fo0bar> 32bit ARCH=ppc compiles, but silently drops about a quarter of your oldconfig settings
02:09<fo0bar>23:29 <@Octal> poor fo0bar
02:09<fo0bar>they're trying to consolidate 32bit and 64bit powerpc source trees into one tree
02:11<fo0bar>caker: do you have an ARCH=um config file yet?
02:11<@caker>not for 2.6.16
02:11<@mikegrb>fo0bar: caker doesn't do uml
02:12[~]caker hopes the break the 4-min kernel compile mile
02:12<fo0bar>if I do 2.6.16, I'll probably want to upgrade the UML demo guest as well... but I'm feeling particularly lazy to wade through the oldconfig changes
02:12<fo0bar>caker: hah!
02:12<@caker>and ...
02:13<@caker>real 3m8.766s
02:13<@caker>user 8m27.320s
02:13<@caker>sys 1m15.980s
02:13<fo0bar>my kernel builds take about 1-1.5 hours each
02:13<@caker>^-- Xenode
02:13<fo0bar>but that's because it's EVERYTHING, minus sound and video modules
02:13<@caker>2.6.16, make oldconfig; time make -j6
02:13<@caker>er make defconfig
02:13<@caker>all that was after a distclean, too
02:13<@caker>ok, and for UML...
02:13<fo0bar>and on a 2.4ghz celery and a 1.25ghz mac mini, respectively
02:14<fo0bar>and not UML
02:15<fo0bar>2.6.16 is funny... the mini booted and cpuinfo showed something like 1249.666666666666666667MHz
02:16<@caker>I read something about that .. perhaps LKML .. /proc/cpuinfo values are messed up
02:16<@caker>something to do with cpufreq
02:16<fo0bar>also, you may find this funny: if you boot, say, a 64MB RAM UML guest on a machine where TMPDIR doesn't have 64MB free, you get OOPSes when you eventually reach the nonexistent memory
02:17<@caker>a host oops?
02:17[~]mikegrb reaches the nonexistant fo0bar
02:18<@caker>also, sometihng new in UML (not sure if it's in 2.6.16) is hotpluggable memory, and it does Bad Things if TMPDIR isn't tmpfs
02:18<@mikegrb>caker: you should get a cd they are super pretty
02:18<@caker>mikegrb: I got one for Christmas
02:18<fo0bar>mikegrb: I didn't give you one because I hate you
02:18<@mikegrb>did it look like that picture?
02:18<fo0bar>you had to win it in a game of chance
02:18<@caker>mikegrb: yes
02:18<@mikegrb>fo0bar: I hate you too
02:19<fo0bar>caker got one because of his kickass help with the UML functionality in finnix
02:19<fo0bar>again: caker = helpful. mikegrb = lucky.
02:20<fo0bar>speaking of which, giving away 1 CD per week is actually saving me money
02:20<fo0bar>I no longer get gummed up inkjet nozzles
02:20<@mikegrb>see, I am helpful
02:20<fo0bar>you're recommended to print on these printers at least once every 2 weeks
02:21<fo0bar>and legit Finnix orders are sporadic (mostly right after a release)
02:21<fo0bar>so sometimes I'll go weeks without printing one
02:22<fo0bar>and the only solution is wasting 3/4 of a set of cartridges running the "clean nozzles/diagnostics" thingy
02:22<fo0bar>and those cartridges are rather expensive
02:23<@caker>I spent $55 bucks on cartridges last week .. I think I only spent $150 on the printer
02:23<@mikegrb>har you think that's bad
02:23<@mikegrb>ours are $80 a piece and it takes 4!
02:23<fo0bar>$100 for the printer here, $80 for a set of cartridges directly from epson, or $20 from a knockoff manufacturer
02:24<@mikegrb>but the pictures are oh so pritty
02:24<fo0bar>(quality is the same though, AFAICT)
02:24<@mikegrb>and we didn't get the $150 mail in rebate because they said our rebate that was sent four weeks before the deadline was post marked after the deadline :<
02:25<fo0bar>meanwile, I'm still using the same toner cartridge that came with the laserjet 1200 I bought 4 years ago
02:25<@mikegrb>same here but it isn't nearly as old
02:25<@caker>I'm all for laserprinters for anything business related/only b&w needed .. inkjet is just too expensive
02:25<fo0bar>$300 printer, but has already paid for itself many times over
02:25<@mikegrb>caker: our laser printer is color!
02:25<@mikegrb>heidi prints her business cards
02:26<@mikegrb>and labels for samples and such
02:26<fo0bar>the quality of color laser has never impressed me
02:26<fo0bar>we have a color Ricoh at work that cost $9 billion or something... OK printouts, not great
02:26<@mikegrb>it says we are at 46% black toner and =~ 76% for the others
02:27<@mikegrb>fo0bar: I've been pretty happy with ours
02:27<@mikegrb>heidi's business cards look just as good as the ones she had printed before
02:27<@mikegrb>and the few pictures I've printed have come out great
02:27<fo0bar>Well, I'm a sucker for raised ink business cards
02:27<@mikegrb>oh and sunday dilbert comics are superb
02:27<fo0bar>but those cost a lot
02:28<@mikegrb>she had full color so not raised ink
02:28<fo0bar>do you have any old business cards?
02:28<fo0bar>I collect them and make little cubes out of them
02:29<@caker>I think I have a box of 5000 or something stupid
02:29<@caker>I used about 20 of them
02:29<@caker>(previous employment)
02:30<fo0bar>they're pretty cool when you get 20 cubes together in a level 1 sponge
02:30<fo0bar>I don't yet have the quantity (or time) to make a level 2
02:32<fo0bar>caker: do you know anything about vserver?
02:32<@caker>fo0bar: enough that I avoided using it. I follow the mailing list .. that's about it
02:33<fo0bar>I don't know much about it, except it's something like freebsd's jail functionality (which I also know little about)... I just noticed debian unstable now has -vserver kernel builds
02:34<fo0bar>and I'm thinking "typical debian... uml is widely used, xen is becoming very popular, and debian choses this other guy"
02:35<fo0bar>ironically, make-kpkg (debian's automated kernel build system) has support for -um, -xen0 and -xenU builds, but it's never been used for debian itself
02:35<fo0bar>I take it back, there are UML kernel packages, but they are always horribly out of date
02:38<@caker>well, the unstable one is recent
02:38<@caker>I/we used to field SO many of the questions in #uml from debian users trying to use ancient/broken UML packages
02:38<fo0bar>but when the functionality is built into the kernel now, you would think they would just treat it as another arch
02:39<@caker>I think they all read in a mag somewhere "apt-get install uml" or whatever
02:39<@caker>real 15m36.277s
02:39<@caker>user 11m4.890s
02:39<@caker>sys 4m22.970s
02:39<@caker>^-- uml compile of the same kernel
02:40<fo0bar> <-- previous debian kernel releases with SKAS added
02:40<@mikegrb>3 straights in a row
02:40<fo0bar>caker: those are both on idle systems?
02:40<@mikegrb>pitty it isn't real money
02:41<@caker>fo0bar: no, the UML is on a production box, but I watched top, there was idle time
02:44[~]mikegrb goes to bed
02:44<fo0bar>caker: anyways, before I go to bed, I'll start uploading today's snapshots. care to try them tomorrow if you have time?
02:44<@caker>fo0bar: in UML, sure ... I still have to implement Finnix for Xenode, but you're next in line
02:44<fo0bar>I was thinking the CD, but UML works too :)
02:45<@caker>oh, sure
02:45<@caker>either way
02:45<@caker>I have two boxes here, amd and xeon
02:46<fo0bar>ok, I'm off to the store to get a couple cases of soda (I've literally be cafeineless for 8 HOURS now), then off to bed
02:46<fo0bar>oh, dell 20.1" widescreen == awesome
02:46<fo0bar>as I'm typing on it
02:47<@caker>dell 24" widescreen == awesome as well :)
02:47|-|avid [] has quit [Ping timeout: 480 seconds]
02:48<TheFirst>you two and your lcds suck!
02:50|-|sysdivision [] has joined #linode
02:50<sysdivision>hi there
02:50<sysdivision>is there any linode operator here?
02:50<TheFirst>as of 3m ago there was
02:51<sysdivision>oh shit
02:55<linbot>New news from forums: Xen Performance Teaser in Xen Public Beta <>
02:56|-|sysdivision [] has quit [Quit: CGI:IRC (Ping timeout)]
03:40|-|myfnp [] has quit [Quit: CGI:IRC (Session timeout)]
03:49|-|fake [] has quit [Remote host closed the connection]
04:28|-|myfnp2 [] has quit [Quit: CGI:IRC (Session timeout)]
04:41|-|avid [] has joined #linode
05:43|-|schultmc [] has quit [Read error: Connection reset by peer]
05:45|-|schultmc [] has joined #linode
06:24|-|linville [] has joined #linode
07:06|-|fake [] has joined #linode
07:34<encode>ok, how do i get my network card (realtek rtl8139) to run at 100mbps full duplex?
07:34<encode>dmesg says "eth0: link up, 10Mbps, half duplex"
07:46<cmantito>ifconfig eth0 up netmask media 100baseT ?
07:48<cmantito>encode: that might do it, but I'm not positive.
07:49<encode>mii-tool -v --advertise=100baseTx-FD
07:49<encode>that does the trick
07:49<cmantito>ah, k then.
07:49<encode>im just not sure why the card doesnt negotiate for 100 full to begin with
07:49<cmantito>it should.
07:49<cmantito>odd switch maybe?
07:50<cmantito>or low quality cable?
07:50<encode>the switch only shows it at 10 half from post all the way through till that line that i've added in /etc/rc.local
07:50<encode>no idea, i guess i can try another cable
07:50<encode>on the bright side, dapper seems to cope well with the rocketraid 100 i recently bought
07:53<encode>the thing that really bugs me is the builtin nic (SiS900) doesnt seem to work - dmesg shows it having a problem reading the mac address and then bailing
07:56<cmantito>hmm.. is there a way to tell the last time an X session running as another user was active?
07:56<encode>the timestamp on their $HOME/.Xsession.<something i dont remember> ?
07:56<cmantito>wonder if I have access...
07:57<encode>(just guessing here)
07:57<cmantito>soudns right, but I ain't got access.
07:57<cmantito>oh well.
08:01[~]encode goes to bed
08:06|-|FireSlash [] has joined #linode
08:26|-|Redgore [~Redgore@] has joined #linode
08:36|-|aster [] has joined #linode
08:42<iggy>caker: my boss chickened out on the 3 year deal, but we are still going to get 2 for a year to start off with... any guesstimates on when 320's will be available?
09:11|-|jekil [] has joined #linode
10:06|-|Dreamer3 [] has quit [Ping timeout: 480 seconds]
10:10|-|Dreamer3 [] has joined #linode
10:26<warewolf>my $node = defined($ARGS{node}) ? $ARGS{node} : int( (scalar @_) /2 ) == scalar @_ /2 && scalar @_ != 0 ? $_[0] : "WikiHome";
10:27<efudd>heah. awesome.
10:27<efudd>that is *GREAT*.
10:27<warewolf>one ?: inside another is painful.
10:27<warewolf>and yes I just wrote that.
10:27<efudd>oh. that's yours? haheahah.
10:27<efudd>Evil Bastard.
10:27<warewolf>It Works(tm)
10:27<efudd>I love ternary.
10:28<warewolf>yeah it has it's uses
10:28<warewolf>the above, an elegant one.
10:28<warewolf>painful, but elegant.
10:28<efudd>So.. you are writing an instance-able wiki...
10:29<warewolf>no that was detecting what node to pull up at request time
10:29<efudd>oh. I was referencing "node" as "BBS Node"
10:29<efudd>vs. "page node"
10:29<warewolf>did you see the last of the line?
10:29[~]warewolf fwhaps efudd
10:29<efudd>Mind in wrong place this AM.
10:29<warewolf>oh dude I've been working 14 hours now.
10:29<efudd>push eax, eax;
10:29<efudd>^^ me
10:30<efudd>trying to do a core analysis
10:30<warewolf>KEEP IT AWAY
10:30<efudd>i'm sucky at it. (these are 300+megabyte images)
10:30<warewolf>you're decompiling a core dump?
10:30<warewolf>please tell me no
10:31<efudd>isn't that what someone does with a core? :)
10:31<warewolf>you're a brave man.
10:31<efudd>tryign to entertain myself @ work. yes.
10:31<warewolf>well, core could mean many things
10:31<warewolf>a core dump, or a DSP core
10:31<efudd>system image at A Time.
10:31<warewolf>if you're an ASM hacker, you should go figure out the instruction set on the BCM43xx (broadcom wireless) cores.
10:32<warewolf>Yes! Reverse engineering at it's finest!
10:32<efudd>i deleted a code set for that chipset last week in fact.
10:32<efudd>somoene had been working on it...
10:32<efudd>i found a codeset a few months back where someone was trying to get that chipset working natively under linsux.
10:32<warewolf>what do you mean you deleted code for the broadcom chipset
10:32<warewolf>efudd- it's working.
10:32<efudd>i sold the laptop, so i deleted the code images that I found ThoseMonthsAgo.
10:33<warewolf>it's not "trying to get working" it's "got working and now fixing bugs and supporting more chipsets"
10:33<efudd>ah, don't need ndiswrapper stuff anymore?
10:33<warewolf>ndiswrapper blows goats
10:33<efudd>tell me about it.
10:33<warewolf>kismet works on supported bcm43xx cards :)
10:33<efudd>i tried it once, got it working, went "neat" and never touched the wireless card again.
10:33<warewolf>go hit that page up
10:34[~]efudd just got his ida pro 5.0 upgrade
10:34<warewolf>efudd- so that ternary operator mess I pasted earlier?
10:34<warewolf>here's the comment (for myself the next time I hack this code)
10:34<warewolf># THAR BE DRAGONS HERE
10:34<efudd>11:34 <@tober> although I personally object to int( (scalar @_) /2) == scalar @_/2, that's pretty bush-league. Something like ( (int(scalar @_))&1)==0 would be better. Much better.
10:35<efudd>fix it!
10:35<efudd>tober is right. :)
10:35<warewolf>er oh
10:35<warewolf>that is even better
10:35[~]efudd nod
10:35<efudd>cleaner too.
10:35<warewolf>I don't do bifting shits enough
10:35<warewolf>that's why I don't do that shit
10:37<warewolf>so what channel is tober in?
10:37<warewolf>I assume some perl channel
10:37<efudd>oh. no.
10:37<efudd>tober is, er, elsewhere.
10:38<efudd>tober does stuff like, hack gcc.
10:38<warewolf>oh my damn.
10:38|-|shad [~Dimedrol@] has joined #linode
10:38<warewolf>pass along a "Thanks!" for the suggestion for me?
10:39<shad>hi all
10:39<warewolf>good morning, afternoon, or evening, whichever the case may be
10:40<shad>At us in Russia evening
10:40<warewolf>I work over-night so it's effectivly evening for me right now, eventhough it's 11:40 AM here :)
10:41<shad>At me the bad translator but at us 11:40 PM
10:42<bendy24>other side of the world!
10:43<shad>Yes a difference in 12 hours
10:44|-|macdan [] has quit [Remote host closed the connection]
10:44<shad>You have woken up and we to sleep
10:44<kvandivo>based on the typical email that i get from russia, I can only assume that you are surrounded by dozens of voluptuous women who are looking for men.
10:44|-|macdan [] has joined #linode
10:45<bendy24>shad: kvandivo wants to order a russian bride
10:46<shad>Russian for Russian
10:47<bendy24>what happens when all the hot russians have moved to north america?
10:47<shad>I with your girls would have fun
10:48<warewolf>this has to be the most interesting conversation of all time
10:48<shad>Smoke a grass?
10:48|-|macdan [] has quit [Remote host closed the connection]
10:49<warewolf>shad- you are here seeking assistance?
10:50<shad>Is not present simply boringly want to communicate to people from other continent
10:50<shad>You normally understand me?
10:51<bendy24>all your bases are belong to us
10:51<warewolf>shad- I mean no offense, but this channel is assistance for customers of
10:52<warewolf>shad- The #moocows channel is better to communicate openly with no defined topic for discussion
10:53<shad>You expel me?
10:53|-|FireSlash [] has quit [Read error: Connection reset by peer]
10:53<warewolf>no, I do not expel you.
10:53<warewolf>you are welcome to stay!
10:54<efudd>s/\!/\ on topic\!/
10:54<alnr>wow escapes and everything
10:54<warewolf>! does not require escapes in PCRE
10:55<alnr>especially on off topic warnings
10:56<shad>Than be engaged?
10:56<efudd>warewolf, pedanticness.
10:56<shad>Thaw, study?
10:56<warewolf>s/\t\h\e\n\ \w\h\y\ \n\o\t\ \d\o\ \w\h\a
10:57<warewolf>+\t\ \q\u\o\t\e\m\e\t\a\ \d\o\e\s\ \i\n\ \p\e\r\l\?//;
10:58<bendy24>oh well
10:58<bendy24>LUNCH TIME
10:58<warewolf>laziness, hubris, and impatience.
10:58<shad>Boringly with you
10:59<warewolf>wow I'm mustingly difficultly finding boringness of person this annoying quite.
10:59<warewolf>try that on for size, translator.
11:00<shad>Manner of dialogue different
11:01[~]warewolf laughs
11:01<alnr>not anymore
11:01<@caker>in my world
11:01<warewolf>shad- you will experience conversation enjoyment in #moocows.
11:02<shad>There there is nothing to do(make)
11:03|-|shad [~Dimedrol@] has left #linode [sans phrases ]
11:04<warewolf>I know #linode is logged
11:04<warewolf>and this will make me look like an asshole
11:05<warewolf>that convo is also being blogged for humor.
11:05<warewolf>"Manner of dialogue different" was great
11:06<warewolf>is it bad that I was filled with glee that I broke his translator? 8-)
11:08<warewolf>suprisingly enough, babelfish translated my attempt at politely telling the guy to go join #moocows from english to russian back to english intact
11:08<kvandivo>you didn't notice the <INPUT TYPE="HIDDEN" NAME="ORIGINALSTRING" VALUE="...."> in the html, huh?
11:09<warewolf>"I will intend no offence, but this channel will be aid for clients"
11:09<warewolf>that's what it translated to eng -> rus -> eng
11:10<warewolf>phrasing things so that it is very difficult to misinterpret is .. more difficult than I thought
11:16|-|afv-13 [] has joined #linode
11:22<warewolf>when it updates
11:24<warewolf> works too.
11:29<@caker>< bendy24> all your bases are belong to us
11:30<warewolf>it made me feel good to break his translator.
11:30<warewolf>"Manner of dialog different"
11:30<warewolf>gee, ya think?
11:31[~]warewolf goes back to working on home-brew-wiki
11:32[~]iggy stabs pdflib documentation in the throat
11:32<warewolf>ITYM 'FTSO'.
11:33[~]warewolf ROFL
11:34<warewolf>googling for FTSO has wli in #kernelnewbies using it in reference to GOBS of kernel code
11:34<iggy>has something to do with bios/acpi I think
11:34<warewolf>FTSO == "fuck the skull of"
11:35<warewolf>commonly used as a term of extreme displeasure with something
11:35<iggy>it also has something to do with kernel stuffs
11:38<warewolf>oh my damn, that felt good
11:38[~]warewolf just popped a good majority of the vertebre in his back twisting his back around
11:39<warewolf>I might actually be able to sleep w/o any pain (bad bed) today.
11:39<warewolf>g'night folks!
12:03|-|jekil [] has quit [Quit: Leaving]
12:05|-|afv-13_ [] has joined #linode
12:05|-|afv-13 [] has quit [Read error: Connection reset by peer]
12:10|-|afv-13 [] has joined #linode
12:10|-|afv-13 [] has quit [Quit: ]
12:11|-|afv-13 [] has joined #linode
12:11|-|afv-13_ [] has quit [Read error: Connection reset by peer]
12:30<linbot>New news from forums: Random Computer Names in /dev/random <>
12:34|-|macdan [] has joined #linode
12:42<linbot>New news from forums: What's the best way to start using a linode VPS? in Sales Questions and Answers <>
12:53<efudd> holy crap that is funny
12:59|-|afv-13 [] has quit [Ping timeout: 480 seconds]
13:40|-|aster [] has quit [Quit: CGI:IRC (Session timeout)]
13:45|-|afv-13 [] has joined #linode
13:49<gpd>efudd: nice... :heh
13:54|-|spr [] has joined #linode
14:12|-|Sgeo [] has joined #linode
14:41|-|jekil [] has joined #linode
14:43<TheFirst>the one linked in the /. comments about that story is good too....
14:57|-|jekil [] has quit [Read error: Connection reset by peer]
15:09<cmantito>anyone familiar with MySQL's licensing scheme?
15:10<linbot>New news from forums: Announcement: Xen Public Beta in Xen Public Beta <>
15:13<gpd>caker: you didn't mention possible Dallas Xen Beta - is that no go?
15:16[~]caker does a little dance
15:16<@caker>gpd: probably not. We'll be done with it (hopefully) by the time I have machines there
15:16[~]caker moshes with you
15:17<@caker>does that post read OK?
15:19<gpd>reads fine - are you editing it? I keep getting email notifications...
15:19<@caker>I made one edit
15:20<@caker>I ddin't know it did that for each edit .. odd
15:21<gpd>maybe some other problem - I got 3 emails for #9581
15:21<@caker>yeah, actually I think I made two edits, so that's probably right (although crappy)
15:21<gpd>not a problem
15:22[~]caker fires off an email of migration notifications for the first set of xen testers
15:24<@caker>bendy24: get one? :)
15:24<bendy24>"migration should be seamless..."
15:24<gpd>how many ppl willing to switch IP address?
15:25<@caker>gpd: a dozen out of 38 or so
15:28[~]gpd considers reconsidering not considering to move IP address
15:35|-|jekil [] has joined #linode
16:03|-|linville [] has quit [Quit: Leaving]
16:18<iggy>how long does this move take?
16:19<@caker>someone else is migrating too
16:19<@caker>so, I dunno .. another 10-15 minutes?
16:23<@caker>iggy: #linode-xenbeta, please
17:03|-|spr [] has quit [Quit: Spoon!]
17:18|-|Marcel [] has quit [Ping timeout: 480 seconds]
17:52<encode>whats the vi shortcut for going to the beginning of a file?
17:52<encode>i know :$ goes to the end
17:54<Spads>and you don't use a colon in vi for that
17:54<Spads>unless you *really* want to use ex mode
17:54<Spads>just hit 0 or $ in command-mode
17:54<@mikegrb>yes you do
17:54<@mikegrb>if you want to go to the beginning of the file
17:54<Spads>end of file
17:54<Spads>I get it
17:54<Spads>you want gg and G
17:54<@mikegrb>:0 works
17:55<Spads>but that's in ex
17:55<Spads>when people say "in vi" I usually assume they mean actual vi keystrokes, rather than dropping into ex mode
17:55<@mikegrb>G and 1G
17:55<Spads>gg works too
17:56<@mikegrb>when I hear people say "in vi" I usually assume they mean whatever will accomplish $task after typing "vi $filename"
17:56<@mikegrb>since that is normally what they mean
17:56<@mikegrb>especially since he said ":$"
17:57<Spads>I caught that eventually
17:57<Spads>I misread "file" as "line" was all
18:05|-|jekil [] has quit [Remote host closed the connection]
18:15<encode>thanks guys ;)
18:23|-|Newsome [] has joined #linode
18:29|-|TheFirst [] has quit [Ping timeout: 480 seconds]
18:30<gpd>random q: my machine at home dual boots breezy/dapper atm - right now when I ssh into it i have to keep deleting the key from known_hosts - what *should* I be doing?
18:30<ElectricElf>gpd: You have two options
18:30<ElectricElf>gpd: First, copy /etc/ssh/ssh_host_*key* from one machine to the other
18:30<ElectricElf>gpd: (Er, from Breezy ot Dapper or vice-versa)
18:30<ElectricElf>gpd: That'll give 'em the same host key. Of course that could be considered a vulnerability.
18:31<ElectricElf>gpd: The other option depends on the clientside SSH version you're using
18:31<ElectricElf>gpd: If you're using a recent enough version of OpenSSH on the clientside, you can specify 'HostKeyAlias', which will look the machine up in ~/.ssh/known_hosts using a different key
18:31|-|TheFirst [] has joined #linode
18:32<ElectricElf>gpd: That requires that you provide an option to ssh each time you connect to one of the machines (ie: '-o HostKeyAlias=foo'), or connect to them via different names (in ~/.ssh/config, using HostName and HostKeyAlias to make them go to the right real hostname and use the appropriate HostKeyAlias)
18:34<gpd>ElectricElf: thanks! I might do your first option - security isn't a major concern on that machine
18:37<gpd>hmm Dapper default has PermitRootLogin yes :!
18:38<ElectricElf>So? :)
18:38<Spads>gpd: file a bug
18:38<gpd>Isn't that a bit dodgy? esp with all the bruteforce attacks atm
18:38<ElectricElf>There wasn't even a debate over that
18:38<Spads>I mean
18:38<Spads>hasn't Dapper re-introduced a root account for some reason?
18:39<Spads>or did I hear wrong?
18:39<ElectricElf>(I mean, it's sufficiently non-controvertial that nobody really argued with that default)
18:39<gpd>no - I think Automatix introduces a root account (maybe where you heard)?
18:39<Spads>well then that's okay I guess
18:39<ElectricElf>gpd: The answer to that is "use a decent password", not "cripple most peoples' machines"
18:39<Spads>they need to do the PermitRootLogin keys or whatever that is
18:40<Spads>who the hell still uses PASSWORDS for remote auth?
18:40<Spads>you shouldn't be sshing around as root anyway
18:40<ElectricElf>Here's news - people with shitty root passwords are likely vulnerable in ways much more convenient than brute-force attacks :)
18:40<ElectricElf>Spads: Exactly - the non-argument goes that if you actually need to, you really fucking need to :)
18:40<gpd>Spads: ssh as root is useful for rdiff-backup (ducks)
18:40<Spads>ElectricElf: that's wonderful logic. Let me punch you in the face now and point out that your gut was far more vulnerable anyway
18:41<Spads>gpd: I use it for just that, and I have PermitRootLogin set to use KEYS ONLY
18:41<Spads>[nick@frotz(~)] grep PermitRoot /etc/ssh/sshd_config
18:41<Spads>PermitRootLogin forced-commands-only
18:41<Spads>that's what it is
18:41<ElectricElf>Finding a default PermitRootLogin set to no can be a real bloody spoilsport when the admin has accidentally locked themselves out of their semi-privileged account (assuming they decided that having a root password set was a good backup plan)
18:41<Spads>works GREAT
18:42<Spads>ElectricElf: I'm very sorry to hear about your problem. Why don't you tape your keys to the hood of your car in case you lock your other set inside
18:42<Spads>ElectricElf: I'm sure the carjackers will understand
18:42<Spads>and leave you alone
18:42<ElectricElf>Spads: I have a set locked away in a safe
18:42<ElectricElf>Spads: (Roughly equivalent to having a long randomly-generated root password)
18:42<Spads>enough with this bullshit about *passwords*
18:43<ElectricElf>Spads: Leaving them on the hood would be roughly equivalent to having a password which can be brutre-forced easily
18:43<Spads>use RSA or DSA keys, or opie or something
18:43<ElectricElf>I use OPIE
18:43<Spads>this "password" shit is so 1990s
18:43<Spads>in that case what you mean is "passphrase"
18:43<gpd>!acronym opie
18:43<ElectricElf>I also run production servers, and if there's no statistical chance of compromise, I use whatever facilities are available to ensure that the machine can be maintained in emergency situations
18:43<linbot>gpd: No definitions found.
18:44<Spads>gpd: One-time Passwords In Everything
18:44<Spads>it's an OTP system, and it rocks
18:44<ElectricElf>Spads: Actually in that case I mean a one-time MD5 hash generated from the combination of the seed and my passphrase, iterated over some number of times in relation to the sequence :)
18:44<Spads>I have a javurscript md5 key calculator on my server, for folks to use to get in
18:44<Spads>ElectricElf: yes you do
18:44<Spads>anyway, I have "passwords" shut off entirely
18:45<ElectricElf>gpd: OPIE is great. If your cellphone is J2ME-capable and you can get a jar or whatever on there, OTPGen is wonderful
18:45<Spads>my system doesn't allow them
18:45<Spads>except from console
18:45<Spads>and then only one account has one
18:45<ElectricElf>Spads: You know of any hardware S/KEY/OPIE generators?
18:45<ElectricElf>Spads: With a little numeric keypad?
18:45<ElectricElf>Spads: Something tamper-obvious?
18:45<ElectricElf>Ah hell, I'd be happy with just something. :)
18:46<ElectricElf>The cellphone option is good, but something which can be attached to a keychain would be a nice option
18:47<gpd> maybe somthing in there?
18:47<ElectricElf>gpd: Those are all software generators
18:48<gpd>I have no clue... and that website doesn't exactly help explain
18:49<ElectricElf>gpd: S/KEY might be a better search term
18:49<ElectricElf>gpd: OPIE is an S/KEY implementation; there are trademarks with that though
18:49<Spads>I worked for a while at VA
18:49<Spads>and they insisted at one point that everyone get a palm pilot
18:49<Spads>with a generator on it
18:49<ElectricElf>Typical of VA :)
18:50<Spads>since they didn't want it to be a net-connected machine
18:50<Spads>and I said no I wanted a piece of paper with like 500 passwords on it
18:50<Spads>and I'd just use that
18:50<ElectricElf>Yeah, and that obviously prevents the sniffing vector
18:50<Spads>and sure enough, there'd be times in the field when a colleague's pilot would be grabbed at a security checkpoint
18:50<ElectricElf>Doesn't prevent the over-the-shoulder vector though
18:50<Spads>and my paper came through
18:50<Spads>but the sysadmins got FURIOUS
18:50<Spads>blah blah blah
18:51<Spads>note that it was ust a page of raw unadorned phrases
18:51<Spads>you know, the six-word hashes
18:51<ElectricElf>Yeah. At that point it's not much better than an RSA keyfob
18:51<Spads>no indication of machine or account or anything
18:51<Spads>they're still mad, but I still say MY PAPER WAS NEVER HOTSYNCED LIKE THOSE DAMN PILOTS
18:51<Spads>and also my paper didn't run proprietary software
18:51<ElectricElf>Yeah, I don't trust a Pilot for that sort of thing
18:52<Spads>which is pretty much anathema to security. period.
18:52<Spads>but marc merlin is still mad at me for being a free software snob about those pilots
18:52<ElectricElf>Spads: Did the generator software support saving the passphrase in an unencrypted manner?
18:52<ElectricElf>Spads: (Or did they use Strip?)
18:52<Spads>I think you had to pen in your passphrase each time
18:52<ElectricElf>If such wasn't feasible, then no worse than paper
18:52|-|FireSlash [] has joined #linode
18:52<ElectricElf>Spads: No worse than paper then really
18:53<Spads>which led peopl eto choose easily-graffiti'd phrases
18:53<Spads>whereas mine was a whole paragraph from something
18:53<ElectricElf>hardware_fob++ :)
18:53<Spads>because I knew I'd nevr have to type it
19:13|-|spr [] has joined #linode
19:18|-|tompuppy [] has joined #linode
19:18<tompuppy>!seen caker
19:18<linbot>tompuppy: caker was last seen in #linode 2 hours, 54 minutes, and 59 seconds ago: <caker> iggy: #linode-xenbeta, please
19:18<@caker>tompuppy: hello
19:19<tompuppy>i just want to know if there is any progress on xen public beta
19:19<tompuppy>i have signed up the beta with the account "acot"
19:20[~]caker checks
19:21<@caker>tompuppy: check your email -- your migration has been configured :)
19:24<warewolf>$ tail -f /var/log/httpd/ | perl -nple '$|=1; $_ =~ s/\\n/\n/g'
19:24<warewolf>holy shit is that useful
19:26<@caker>tompuppy: you should /join #linode-xenbeta
19:28<efudd>warewolf, er, eh? I'd have to look up $| i guess.
19:28<efudd>(and '-p' for that matter.)
19:28<fo0bar>caker: make with the finnix xen :)
19:30<@caker>fo0bar: hmm?
19:30<fo0bar>it wasn't funny :)
19:31<warewolf>efudd- my HTML::Mason environment instead of dumping a HTML::Mason error to the web browser, dumps it to a _single line_ in apache's error log
19:31<warewolf>efudd- so what that does, is it changes the '\n' literal strings back into newline characters.
19:31<efudd>I was trying to figure out why you might want to do that.. just realized "error_log"
19:32<gpd>tr '\\n' '\n' ?
19:32<warewolf>[Mon Mar 27 20:27:44 2006] [error] Request as string: GET /wiki/?abcd HTTP/1.1\nAccept: text/xml,application/xml,appl
19:32<warewolf>ication/xhtml+xml,text/html;q=0.9,text/plain;q=0.8,image/png,*/*;q=0.5\nAccept-Charset: ISO-8859-1,utf
19:32<warewolf>with taht one-liner, it becomes MUCH MORE READABLE
19:33<warewolf>efudd- oh and I realised something from our earlier convo
19:33<warewolf>efudd- under HTML::Mason, @_ is always even :(
19:34<efudd>well, no worries about the carry bit then!
19:34<warewolf>efudd- so I had to create a temporary variable to pull mod_perl's $r->args() into, and _THEN_ do that bifting of shits
19:35[~]efudd continues to try to incorporate x86 CPU registers and bits into normal conversation in an attempt at getting them firmly in head
19:35<warewolf>and yes, I like saying bift shiting
19:35<efudd>Ah, well, yes. just push ebp into eax and shr 2 :)
19:35<efudd>don't forge tto store eax somewhere nice before restoring esi to ebp.
19:36<efudd>(that's almost making sense. it will, soon. I don't quite fully grok() BP vs SP vs. $OTHERS as well as i grok() the alphabet.)
19:40<warewolf>that is ... interesting.
19:40<efudd>Wow, the Sarcasm flag in my SSE3 extensions is blinking orange.
19:41<warewolf>my $node = defined($ARGS{node}) ? $ARGS{node} : scalar @args & 1 == 1 ? $_[0] : "WikiHome";
19:42<warewolf>oh, and my @args = $r->args();
19:42<warewolf>above that.
19:42[~]efudd nods
19:43<warewolf>suprisingly enough, that manages to function in a predictable way, when the webserver gets a request string like
19:43<warewolf>... which I wouldn't expect it to process correctly
19:43<warewolf>but it thinks the Wiki node in that situation is "and"
19:44<warewolf>no matter where the positioning of "and" in the string, or it's sorted position
19:44<warewolf>it still ends up being at $_[0]
19:46<warewolf>ooh, I'm wrong :(
20:12<fo0bar> <-- caker
20:13<@caker>I have quite a collection of the really small SATA cables
20:14<@caker>and lotsa headsinks and intel stickers
20:15<fo0bar>I saved about 5 of the intel heatsinks, and since then I've just been throwing them away
20:16<fo0bar>I keep the P4 stickers though
20:16<fo0bar>but to tell you the truth, I've run out of "ironic" places to put them
20:16<efudd>why would you do that.
20:17<@caker>I think I have a cache of the (really old) Apple stickers too .. oh, and the 5.25 read-only stickers .. nice silver ones I put on my motorcylce helmet for extra visibility
20:17<fo0bar>efudd: you mean why not put them on the server?
20:17<efudd>er, no. I mean, end up with so manyu SATA cables.
20:17<efudd>that's .. well, more than one should have. Even a geek.
20:17<efudd>Unless you build computers.. then that's different.
20:17<fo0bar>efudd: I build rather large amounts of servers for work
20:17<efudd>Yes, ok.
20:18<efudd>(quantity, btw. unless i'm sorely mistaken.)
20:18<fo0bar>and these days, the motherboard, the RAID controller, and the drive all come with SATA cables
20:18<efudd>I used to find places to stick Fedex saturday delivery stickers...
20:18<efudd>i had a couple of rolls..
20:18<efudd>and by "find" I mean "not be seen whilst placing"
20:34<linbot>New news from forums: Xen Beta Gotchas and Known Issues in Xen Public Beta <>
20:40<linbot>New news from forums: Sticky: Xen Beta Gotchas and Known Issues in Xen Public Beta <>
20:53[~]gpd receives 4 emails re ^^ ;)
20:56<Sgeo>What's Xen?
21:15|-|ElectricElf [] has quit [Quit: ]
21:21|-|ElectricElf [] has joined #linode
21:29|-|FireSlash_ [] has joined #linode
21:36|-|FireSlash [] has quit [Ping timeout: 480 seconds]
21:38|-|Electric1lf [~dbharris@] has joined #linode
21:38|-|ElectricElf [] has quit [Read error: Connection reset by peer]
21:38|-|afv-13 [] has quit [Quit: leaving]
21:44[~]efudd updates blog:
22:07|-|Dreamer3 [] has quit [Ping timeout: 480 seconds]
22:08|-|Dreamer3 [] has joined #linode
22:10<linbot>New news from forums: Reboot: host40 in System and Network Status <>
22:14<encode>if only i had an ip based kvm...dont you hate it when you change something on a remote system, reboot, cross fingers, and it never comes up again
22:15<encode>now i have to wait till i get home from work to figure out what went wrong
22:16[~]caker stabs host40
22:17<efudd>Sigh. I can't upgrade to Visual Studio 2005 due to something calld "Microsoft SQL Native Client" being installed. I can't UNINSTALL it because it needs the original TEMPORARY DIRECTORY that it installed from...
22:17<efudd>I can't UPGRADE it to unstall it becuase the UPGRADE needs that TEMPORARY DIRECTORY that it installed the OLD VERSION from
22:18<efudd>I can't find the OLD VERSION because, well, it isn't available anymore.
22:19<encode>does it say what the original temporary directory its looking for is?
22:20<@caker>that's just silly
22:20<efudd>sure. it says e:\randomshithere\setup\
22:20<efudd>unfortunately, that was a temporary drive
22:20<efudd>which was long ago removed and placed on another machine.
22:20<encode>so net use e: \\localhost\c$
22:21<encode>and recreate the folder structure
22:21<efudd>the "automated" cleanup tools do nothing more than troll the registry and find where they were supposed to be uninstalled from
22:21<efudd>then calls those uninstallers
22:21<efudd>encode, yes, great. the directory STRUCTURE isn't the problem
22:21<efudd>it's the *UNINSTALL SCRIPT* that it can't find.
22:21<encode>yeah, i understand that now
22:21<efudd>because the ASSHAT that created the fucking install didn't think to SAVE IT
22:21<efudd>so now my development machine is RUINED
22:21<efudd>and has to be reinstalled
22:21<efudd>unless i can find a way to clean up 2003+ beta 2005
22:22<encode>what is "microsoft SQL Native Client" installed with? a previous version of visual studio?
22:22<efudd>good fucking question.....
22:22<efudd>Statistics since 10/27/2005 6:34 PM
22:22<efudd>i don't want to reboot this machine. much less reinstall it :)
22:22<efudd>a Windows XP laptop, up since october of 2005 is gold! :P
22:23<efudd>this is my primary gaming and dev box. sigh sigh.
22:23<efudd>oh even better. the automated uninstaller gives me an XML error
22:23<efudd>followed by borked postprocessed html
22:23<efudd>   <!-- div id="RPCreated" style="display:none">
22:24[~]encode makes a mental note never to install visual studio
22:24<efudd>the problem here is that i installed beta 2005 i think
22:27|-|Sgeo [] has quit [Quit: Konversation terminated!]
22:44|-|fake [] has quit [Ping timeout: 480 seconds]
22:47|-|FireSlash_ [] has quit [Quit: Leaving]
22:48|-|fake [] has joined #linode
22:52|-|Newsome [] has left #linode [Linux: Now with employee pricing!]
22:55|-|Electric1lf [~dbharris@] has quit [Quit: ]
22:56|-|ElectricElf [] has joined #linode
22:59|-|VS_ChanLog [] has left #linode [Rotating Logs]
22:59|-|VS_ChanLog [] has joined #linode
23:02|-|tompuppy [] has quit [Quit: Leaving]
23:18<efudd>heh. why is the maximum IPs/linode now 2? :)
23:18<@caker>yes, need more?
23:19<efudd>no i'm at 4.
23:19<@caker>oh, missed the why part
23:19<efudd>just wondering :)
23:19<@caker>rdns kiddiez
23:19[~]efudd nod
23:19<efudd>just perusing through options out of boredom and noticed LPM yelling at me for being over limit.
23:28|-|Marcel [] has joined #linode
23:28<warewolf>I have two and it won't let me have any more :)
23:28<warewolf>one for IRC, one for everyting else.
23:28<warewolf>(yes I'm that paranoid)
23:28|-|Marcel [] has quit [Quit: ]
23:29|-|Marcel [] has joined #linode
23:31<efudd>I'm now getting an insane number of hits from china, korea, etc on my webserver...
23:32<efudd>well, by "insane" I mean considerably more than normal. one every few minutes (~3-5) hitting a non-published site with GET / and nothing else.
23:32<warewolf> probally indexed it
23:32<@caker>warewolf: ready to rock the xen, or what?
23:32<efudd>It'd be odd for someone to index which is Options -Indexes
23:32<efudd>I've changed DNS to redirec tit to localhost for now.
23:33<efudd>now spain has joined in.
23:39<warewolf>caker- can I have another day or two?
23:39<@caker>warewolf: of course
23:40<warewolf>caker- cool. I did set my DNS zone TTLs down to 30 minutes, so hopefully I won't have interesting issues. But I do host websites for a few people and wanted to be sure that they are aware of some down time
23:40<@caker>warewolf: how much used space in your filesystems, total (for resizing down before migration to reduce cross datacenter xfer time)?
23:40<warewolf>oh dear lord
23:41<warewolf>I have only a couple hundred megs free
23:41<warewolf>but I can probally blow away logs and web junk I don't need
23:41<warewolf>(eg, mp3s)
23:41<@caker>ok .. I've been seeing 1.5-2.5 MB/sec between HE<->TP
23:41<warewolf>I hope you don't xfer swap partitions
23:42<warewolf>#$ du -sh .
23:42<warewolf>849M .
23:42<warewolf>I didn't think I had that much web shit online
23:42<warewolf>#$ du -sh .
23:42<warewolf>163M .
23:42<warewolf>there's a chunk
23:42<@caker>I do, since I can see people setting those as swap types in the LPM, but then formatting them something else
23:43<@caker>typically they're small, anyhow
23:43<efudd>sorry, that's funny.
23:43<@caker>sad, but true
23:43<efudd>what, you give us free swap or something? :)
23:44<@caker>not to get around anything, .. just I have no clue what users will do, and it's their data
23:44<efudd>(some of the providers do; so that would explain a tendancy towards folk putting a FS on them.)
23:44<efudd>that'd make me.... well... trash them randomly.
23:44<efudd>for entertainment.
23:44<@caker>forced size swap is useless anyhow ...
23:44<efudd>(there is a reason I'm not a multi-millionaire. I'd be EVIL)
23:44<@caker>considering the dd; swapon /some/file trickery
23:44<warewolf>jesus christ
23:45<warewolf>I have .. oh.
23:45<warewolf>I have 400 megs of BOOKS on my website.
23:45<warewolf>100 megs of mp3s
23:45<efudd>you too? :)
23:45<warewolf>Filesystem Size Used Avail Use% Mounted on
23:45<warewolf> /dev/ubda 4.0G 3.0G 1021M 75% /
23:46<efudd>jason@forever ~/public_html/private $ du -sh books
23:46<efudd>308M books
23:46<efudd>jason@forever ~/public_html/private $ du -sh oreilly_books
23:46<efudd>576M oreilly_books
23:46<efudd>damn. nearly 1GiB.
23:46<Eman>O RLY books?
23:46<warewolf>for using GiB instead of GB or "gig"
23:46<efudd>I figured in $this context, it was needed.
23:47[~]efudd ponders the long term effects of 50mg of diphenhydramine HCL daily is
23:49[~]caker tries to figure out where "w" gets its data from
23:49<@caker>Lish users aren't showing up in "w" on these new boxes, but the do in last
23:49<efudd>oh. do you have screen (mis)configured?
23:49<efudd>This window is now logged in.
23:49<@caker>well, they log into a perl shell first
23:49<efudd>I can "hide" my login by doing ^AL
23:50<efudd>screen is responsible for writing the entry.
23:50<efudd>w just reads what is there.
23:50<efudd>or rephrase. screen is responsible for requesting the login entry written. i forget. clearly.
23:50<@caker>ok, so I'll log into Lish w/o screen involved
23:50<@caker>yeah, we'll see
23:50<@caker>poweroff == they appear in w
23:51<warewolf>screen is suid, and therefore dicks with wtmp
23:51<@caker>ok, nabbed my /etc/screenrc from da other host
23:52<@caker>(that wasn't my uber, all bad keys gone screenrc, that's a diff one)
23:52<efudd>ok. on my second update, I was trying to think about what process would be receiving screens request and fulfilling it. :)
23:52<efudd>suid is clearly the answer.
23:52<efudd>I used to be a damned good SA. Honest. :P
23:53[~]warewolf kicks up2date
23:53[~]warewolf su's to root and runs up2date-nox -u
23:54[~]warewolf runs again, now that up2date has updated itself to date
23:54<efudd>do i have to preach the failures of up2date?
23:54<efudd>Specifically it's heritage?
23:55<warewolf>that's ok
23:56<@caker> -l and -ln
23:56<@caker> turns login mode on or off (for /var/run/utmp updating). This can also be defined through the "deflogin"
23:56<@caker> .screenrc command.
23:56[~]efudd figured it was going to be a no and didn't bother actually removing the soap box from the storage closet
23:57<warewolf>caker- do you compress people's filesystems when you xfer them?
23:57<warewolf>caker- what I'll do is before I actually do the migration is I'll zero out my swap fs
23:58<warewolf>caker- or delete it in LPM and re-create it, which I assume would almost be the same as dd if=/dev/zero of=/dev/swap_device
23:58<@caker>warewolf: yeah, that won't help -- you'll need to use the resize facility before and after migrating
23:58<warewolf>caker- for swap?
23:59<warewolf>caker- I'll take care of shrinking my / fs, don't worry about that
23:59<@caker>warewolf: nod ... I doubt it'll make any diff with the swap, but go ahead
23:59<warewolf>caker- I'm just talking about stupid easy quick fixes, like swap
---Logclosed Tue Mar 28 00:00:11 2006