Back to Home / #uml / 2007 / 09 / Prev Day | Next Day
#uml IRC Logs for 2007-09-14

---Logopened Fri Sep 14 00:00:58 2007
00:05|-|aroscha [~aroscha@chello213047053193.30.11.tuwien.teleweb.at] has joined #uml
00:24|-|force [~force@ppp121-45-10-139.lns10.adl2.internode.on.net] has quit []
01:27|-|aroscha [~aroscha@chello213047053193.30.11.tuwien.teleweb.at] has quit [Quit: aroscha]
02:57|-|krau [~cktakahas@81.253.76.177] has joined #uml
03:13|-|tyler [~tyler@ARennes-256-1-124-96.w90-32.abo.wanadoo.fr] has joined #uml
03:14|-|krau [~cktakahas@81.253.76.177] has quit [Ping timeout: 480 seconds]
03:48|-|peterz [~peterz@i55087.upc-i.chello.nl] has quit [Ping timeout: 480 seconds]
03:58|-|aroscha [~aroscha@chello062178175192.15.14.tuwien.teleweb.at] has joined #uml
04:07|-|peterz [~peterz@i55087.upc-i.chello.nl] has joined #uml
04:11|-|LemyDange1 [~christoph@p5B24FB6E.dip.t-dialin.net] has joined #uml
04:11|-|LemyDange1 [~christoph@p5B24FB6E.dip.t-dialin.net] has left #uml []
04:31|-|tyler [~tyler@ARennes-256-1-124-96.w90-32.abo.wanadoo.fr] has quit [Remote host closed the connection]
04:34|-|tyler [~tyler@ARennes-256-1-124-96.w90-32.abo.wanadoo.fr] has joined #uml
05:13|-|krau [~cktakahas@32-111.252-81.static-ip.oleane.fr] has joined #uml
06:50|-|tyler [~tyler@ARennes-256-1-124-96.w90-32.abo.wanadoo.fr] has quit [Ping timeout: 480 seconds]
07:02|-|tyler [~tyler@ARennes-256-1-19-108.w90-32.abo.wanadoo.fr] has joined #uml
07:04|-|Baltam [~WIKIMOKI@tor-irc.dnsbl.oftc.net] has quit [Ping timeout: 480 seconds]
07:04|-|Baltam [~WIKIMOKI@tor-irc.dnsbl.oftc.net] has joined #uml
07:13|-|aroscha [~aroscha@chello062178175192.15.14.tuwien.teleweb.at] has quit [Quit: aroscha]
07:36|-|balbir [~balbir@122.167.75.103] has joined #uml
07:48|-|aroscha [~aroscha@chello213047053193.30.11.tuwien.teleweb.at] has joined #uml
08:13|-|dang [~dang@aa-redwall.nexthop.com] has joined #uml
08:30|-|krau_ [~cktakahas@32-111.252-81.static-ip.oleane.fr] has joined #uml
08:33|-|krau_ [~cktakahas@32-111.252-81.static-ip.oleane.fr] has quit []
09:33|-|krau [~cktakahas@32-111.252-81.static-ip.oleane.fr] has quit [Quit: Varei!!!]
09:50|-|krau [~cktakahas@32-111.252-81.static-ip.oleane.fr] has joined #uml
09:50|-|jdike [~jdike@pool-71-248-190-161.bstnma.fios.verizon.net] has joined #uml
09:50<jdike>Hi guys
09:50|-|AllenJB [~ajb42@raptor.ukc.ac.uk] has left #uml []
09:51<kokoko1>Hi
10:07<dgraves>morning!
10:07<dgraves>what are we breaking today, jdike?
10:07|-|krau [~cktakahas@32-111.252-81.static-ip.oleane.fr] has quit [Quit: Varei!!!]
10:25<jdike>good question
10:26<jdike>today, I think I will look at caker's network memory allocation failure
10:27<jdike>and look at my genirq code
10:31<dgraves>jdike: you gonna speed that up too? :)
10:31<jdike>speed up everything
10:31<mikegrb>jdike will engage ludicrous speed
10:32<mikegrb>it had been #IFDEFed out
10:32[~]jdike is beyond warp speed now
10:44[~]dgraves hopes jdike doesn't hit a wall or something...
10:45|-|ram [~ram@pool-71-117-233-85.ptldor.fios.verizon.net] has quit [Ping timeout: 480 seconds]
10:45<dgraves>so what's cool and upcoming in the uml space, jdike ?
10:46<jdike>you know about tickless support, right?
10:51<dgraves>vaguely.
10:51<dgraves>we were talking about it the other day.
10:52<dgraves>i put it on my list to figure out why its cool. ;)
10:56<caker>jdike: that would be great, thanks
10:56<jdike>dgraves, ask caker why it's cool :-0
10:57<jdike>probably literally
10:58<caker>jdike: so, this may tie-in to the tickless topic, actually -- we recently sponsor-hosted a programming contest for RoR -- http://railsrumble.org
10:59<jdike>Host railsrumble.org not found: 3(NXDOMAIN)
10:59<caker>about 130 teams, we distributed the accounts across 4 machines. They were able to deploy before the contest started
10:59<caker>sorry, .com
10:59<caker>.. but they weren't allowed to boot. So, naturally, everyone booted a the same time and started setting up their deployment environments (rails stack, tools, blah blah)
11:00<caker>it was fun (not) to see how UML and the host handlded the load
11:00<caker>context switches hit the wall at around 500k/sec
11:01<jdike>tickless might help there
11:01<caker>Nothing crashed, the UMLs chugged along. That lasted for a few hours, but it was painful to watch
11:01<caker>yeah
11:01|-|aroscha [~aroscha@chello213047053193.30.11.tuwien.teleweb.at] has quit [Quit: aroscha]
11:01<jdike>~33 UMLs per host, that's 3300 interrupts/sec the host has to deliver and the UMLs have to react to
11:01<caker>2.6.21 .. I wasn't ready/comfortable with making 2.6.22 the "latest" yet since it hadn't received much testing, and I wasn't sure if there was any significant perf gains to warrant the risk
11:02<jdike>that doesn't really compare to 500K context switches/sec though
11:02<caker>heh. the split was more like 45/46/30/20
11:04<caker>have any idea of how much tickless will translate into real-world performance gains?
11:06<jdike>for a single UML, it gives me 1-2% on a kernel build
11:06<jdike>for a busy host, it should help more
11:07[~]jdike fixes an Ingo-bug
11:07<caker>another data point -- host load on during that time was 10% user, 90% system :)
11:10<jdike>hmm
11:11<jdike>that smacks more of a UML system call load
11:11<caker>Yes
11:11<caker>I was dreaming of UML+kvm
11:11<caker>.. or something ..
11:15<jdike>yeah
11:15<jdike>I've been doing shorter-term things for a bit
11:23<jdike>So tap devices don't do FIONREAD
11:34|-|aroscha [~aroscha@213.129.230.12] has joined #uml
11:37|-|ram [~ram@bi01p1.co.us.ibm.com] has joined #uml
11:39<jdike>plan B is a bit more complicated, but prolly make more sense
12:21|-|tyler [~tyler@ARennes-256-1-19-108.w90-32.abo.wanadoo.fr] has quit [Ping timeout: 480 seconds]
12:24<jdike>Plan B seems to work a bit better
12:49<dgraves>jdike: long as its not plan 9!
12:49|-|JDLSpeedy [~joe@fl-71-52-109-158.dhcp.embarqhsd.net] has quit [Read error: Connection reset by peer]
12:51<jdike>and it's not from outer space
12:54|-|dex [~dex@acnq188.neoplus.adsl.tpnet.pl] has joined #uml
12:56<dgraves>::L::
13:01|-|tyler [~tyler@ARennes-256-1-19-108.w90-32.abo.wanadoo.fr] has joined #uml
13:05|-|JDLSpeedy [~joe@fl-71-52-109-158.dhcp.embarqhsd.net] has joined #uml
13:09|-|mgross [~mgross@jffwprtest.jf.intel.com] has joined #uml
13:21|-|tyler [~tyler@ARennes-256-1-19-108.w90-32.abo.wanadoo.fr] has quit [Remote host closed the connection]
13:28|-|tyler [~tyler@ARennes-256-1-19-108.w90-32.abo.wanadoo.fr] has joined #uml
14:05|-|tyler [~tyler@ARennes-256-1-19-108.w90-32.abo.wanadoo.fr] has quit [Ping timeout: 480 seconds]
14:07|-|balbir [~balbir@122.167.75.103] has quit [Ping timeout: 480 seconds]
14:16|-|dex [~dex@acnq188.neoplus.adsl.tpnet.pl] has quit [Remote host closed the connection]
14:18|-|tyler [~tyler@ARennes-256-1-55-130.w90-32.abo.wanadoo.fr] has joined #uml
14:19|-|aroscha [~aroscha@213.129.230.12] has quit [Quit: aroscha]
14:31|-|balbir [~balbir@122.167.75.112] has joined #uml
15:01|-|tyler [~tyler@ARennes-256-1-55-130.w90-32.abo.wanadoo.fr] has quit [Read error: Connection reset by peer]
15:03|-|tyler [~tyler@ARennes-256-1-55-130.w90-32.abo.wanadoo.fr] has joined #uml
15:40|-|Baltam [~WIKIMOKI@tor-irc.dnsbl.oftc.net] has quit [Remote host closed the connection]
15:45|-|Baltam [~WIKIMOKI@tor-irc.dnsbl.oftc.net] has joined #uml
15:59|-|dang [~dang@aa-redwall.nexthop.com] has quit [Quit: Leaving.]
16:02|-|tyler [~tyler@ARennes-256-1-55-130.w90-32.abo.wanadoo.fr] has quit [Ping timeout: 480 seconds]
16:18|-|tyler [~tyler@ARennes-256-1-75-188.w90-32.abo.wanadoo.fr] has joined #uml
16:52|-|dex [~dex@acnq188.neoplus.adsl.tpnet.pl] has joined #uml
16:55|-|dex [~dex@acnq188.neoplus.adsl.tpnet.pl] has quit []
17:11|-|krau [~cktakahas@81.253.7.89] has joined #uml
17:11|-|tyler [~tyler@ARennes-256-1-75-188.w90-32.abo.wanadoo.fr] has quit [Ping timeout: 480 seconds]
17:15|-|shze [~shze@87.193.58.63] has joined #uml
17:16|-|shze [~shze@87.193.58.63] has quit [Remote host closed the connection]
17:18|-|besonen_mobile [~besonen_m@71-220-234-148.eugn.qwest.net] has quit [Quit: Leaving]
17:23|-|tyler [~tyler@ARennes-256-1-111-123.w90-32.abo.wanadoo.fr] has joined #uml
17:28|-|krau [~cktakahas@81.253.7.89] has quit [Quit: Varei!!!]
17:31|-|mgross [~mgross@jffwprtest.jf.intel.com] has quit [Ping timeout: 480 seconds]
17:39|-|tyler [~tyler@ARennes-256-1-111-123.w90-32.abo.wanadoo.fr] has quit [Quit: ++]
17:53|-|niteowl [~jeff@ita4fw1.itasoftware.com] has joined #uml
17:55<niteowl>I have successfully built uml under x86_64 but it is seg faulting after some of the startup messages appear - I have spent the afternoon googling and poking around the howto and the wiki but must be missing something
17:56<niteowl>Might this be a 2G/2G memory split issue? or would it segfault without any messages in that case
17:56<niteowl>(as an aside grepping around, that CONFIG_HOST_2G_2G) param seems to only be used for ppc uml in it's Makefile - I could find no other reference
17:57<niteowl>(not only that the menuconfig does not show it as an option even for non uml builds)
17:58<niteowl>I'm thinking I may be missing somethings simpler...I'm using a canned root fs I downloaded via a link at the wiki
17:59<niteowl>it gives a warning about not being able to start a console and another about not being able to identify the partition on the root fs - these both seem harmless from what I'm finding via google and the uml wiki
17:59<niteowl>so I'm perplexed to say the least.
18:00<jdike>UML is segfaulting?
18:01<niteowl>it may be my root file system - I have built kernels for 2.6.12, 2.6,17, and 2.6.22 but when I mounted the canned root fs I downloaded, I noticed it's /boot/config indicated that it was 2.6.11 - which might be close enough but maybe not
18:01<niteowl>yes
18:01<jdike>that doesn't matter
18:02<jdike>nothing in /boot matters
18:02<jdike>what UML version?
18:02<niteowl>ok. also, I DID NOT recompile my host kernel - I'm hoping to not need this
18:02<jdike>you don't
18:02<niteowl>what version? Whatever comes from kernel.org in the kernel versions I mentioned
18:02<niteowl>goog
18:02<niteowl>err good
18:03<jdike>so, 2.6.22 is segfaulting, for example
18:03<niteowl>yes
18:03<jdike>is it dumping core?
18:03<niteowl>I didn't look - one moment
18:03<jdike>ulimit -c unlimited
18:03<jdike>and try again
18:04<niteowl>actually looks like a Kernel panic
18:05<jdike>what's the panic?
18:05<niteowl>shall i try the ulimit thing and hope for a core?
18:05<niteowl>Kernel panic - not syncing: Attempted to kill init!
18:05<jdike>the segfault is the result of that - abort is impemented as kill(getpid(), SIGSEGV)
18:05<jdike>yes
18:05<niteowl>ok
18:06<jdike>paste the full boot log someplace like rafb.net
18:08<niteowl>http://rafb.net/p/BGUJpe99.html
18:09<niteowl>I have a number of builds I've been messing around with - I decided to just go back to the canned kernel and canned root fs I downloaded to give you the crash info
18:10<niteowl>the command I used to get you the core: ./kernel64-2.6.22.5 ubd0=root_fs.fc-3-base.x86_64.pristine.20050606.img debug
18:11<jdike>not a lot to go on there
18:11<jdike>where'd you get the filesystem?
18:11<jdike>and is that the only one you've tried
18:11<jdike>?
18:11<niteowl>yes, it's the only one I've tried...let me find the link I used from the wiki
18:15<niteowl>http://www.stearns.org/uml-root/root_fs.fc-3-base.x86_64.pristine.20050606.bz
18:15<niteowl>found it
18:16<jdike>hold on while it arrives
18:16<niteowl>ok
18:20|-|Baltam [~WIKIMOKI@tor-irc.dnsbl.oftc.net] has quit [Remote host closed the connection]
18:22<jdike>now hold on while it uncompresses
18:23|-|Baltam [~WIKIMOKI@tor-irc.dnsbl.oftc.net] has joined #uml
18:24<niteowl>ok
18:26<jdike>for me, it sits there for a bit, then reboots
18:26<niteowl>hmm
18:26<jdike>strange
18:26<niteowl>strange indeed
18:27<jdike>with init=/bin/sh, I get what you're getting
18:27<niteowl>ahh ook
18:27<niteowl>so maybe I should try another root filesystem
18:27<jdike>yeah
18:28<jdike>try something from http://uml.nagafix.co.uk/
18:28<niteowl>ahh ok
18:28<niteowl>brb
18:29<niteowl>they don't seem to have any 64 bit root filesystems
18:29<jdike>all the AMD64 ones are 64-bit
18:29<niteowl>ah ok
18:29<niteowl>will they work on intel though?
18:30<jdike>yes
18:30<jdike>it's all x86_64
18:30<niteowl>ok
18:30<niteowl>i'm downloading fc7 right now
18:33<niteowl>downloaded - need to move it then unpack it
18:35<niteowl>hmm must be tired - let me fetch it again directly to where I want it once second
18:38<niteowl>uncompressing
18:39<jdike>that one looks healthy here
18:39<niteowl>great
18:39<niteowl>uncompress done - testing
18:40<niteowl>looks great!
18:40<niteowl>I have a login prompt
18:40<jdike>root, no password
18:40<niteowl>ok I will need to test that working Root FS on my other machines
18:40<niteowl>with the versions of the kernel I built earlier - but probably not now
18:40<niteowl>Yea! Logged in
18:41<niteowl>shutting it down - works fine too :-)
18:41<niteowl>ok
18:41<niteowl>good very good
18:42<niteowl>It seems there is a bad canned root fs out there unfortunately :(
18:42<jdike>yup
18:42<niteowl>Thank you for your assistance
18:42<jdike>np
18:42<niteowl>I'm glad it was something easy.
18:42<niteowl>Thanks again!
18:42<niteowl>time to go home :-)
18:42<niteowl>see you
19:46|-|besonen_mobile [~besonen_m@71-220-234-148.eugn.qwest.net] has joined #uml
19:54|-|ram [~ram@bi01p1.co.us.ibm.com] has quit [Ping timeout: 480 seconds]
20:31|-|balbir [~balbir@122.167.75.112] has quit [Ping timeout: 480 seconds]
20:43|-|balbir [~balbir@122.167.75.243] has joined #uml
20:58|-|jdike [~jdike@pool-71-248-190-161.bstnma.fios.verizon.net] has quit [Quit: Leaving]
22:51|-|jammaj [jam@deadbeefbabe.org] has joined #uml
22:59|-|VS_ChanLog [~stats@ns.theshore.net] has left #uml [Rotating Logs]
22:59|-|VS_ChanLog [~stats@ns.theshore.net] has joined #uml
23:08|-|_Hunger [Hunger.hu@Hunger.hu] has joined #uml
23:46|-|ram [~ram@pool-71-117-233-85.ptldor.fios.verizon.net] has joined #uml
---Logclosed Sat Sep 15 00:00:11 2007