Back to Home / #xen / 2005 / 04 / Prev Day | Next Day
#xen IRC Logs for 2005-04-07

---Logopened Thu Apr 07 00:00:09 2005
01:16--- <<-- mael [] has quit (Server closed connection)
01:16--- ---> mael [] has joined #xen
01:43--- <<-- knewt [] has quit (Server closed connection)
01:43--- ---> knewt [] has joined #xen
02:46--- <<-- visik7 [] has quit (Read error: Connection reset by peer)
03:03--- ---> jimix [~jimix@] has joined #xen
03:09mael hi
03:10rusty mael: hi!
03:10mael rusty: back in the bush?
03:11muli hey rusty, are you summitting?
03:11rusty mael: well, you could say that 8)
03:11rusty muli: For the Xen Summit? Naah, too much travel for me.
03:12mael so how was your time in europe?
03:12rusty mael: cool (well, cold). Food in Italy was excellent, however.
03:12jimix rusty: muli: its ok you are currently missing pratt vs. drepper
03:13mael you didn't travel in france, did you?
03:13muli jimix, in person?!
03:13jimix in person
03:13muli let me guess, TLS, segmentation, oh my
03:13muli dude, they have working wireless - that's already better than OLS
03:13rusty jimix: wow, drepper is in attendence.
03:14muli jimix, any thoughts on dropping bk by the Cambridge folks?
03:21--- ---> matta-lt [~matta@] has joined #xen
03:34jimix muli: nothing officiall mostly talk in the pubs
03:34jimix muli: wifi was not stable till today..
03:35muli jimix, just like OLS then :-)
03:37--- <<-- jimix [~jimix@] has quit (Remote host closed the connection)
03:37--- ---> jimix [~jimix@] has joined #xen
03:38--- <<-- jimix [~jimix@] has quit (Remote host closed the connection)
03:39--- ---> jimix [~jimix@] has joined #xen
03:45--- <<-- rusty [] has quit (Read error: Connection reset by peer)
03:50--- <<-- jimix [~jimix@] has quit (Ping timeout: 480 seconds)
04:20--- <<-- nextime [] has quit (Quit: No windows for this server)
04:35--- ---> athomas [] has joined #xen
04:49--- ---> Beaky [] has joined #xen
05:31--- ---> nextime [~nextime@] has joined #xen
05:34--- <<-- tab [] has quit (Read error: Operation timed out)
05:36--- <<-- drbyte [] has quit (Ping timeout: 480 seconds)
05:44--- <<-- muli_ [] has quit (Quit: My damn controlling terminal disappeared!)
05:47--- ---> drbyte [] has joined #xen
06:05--- ---> tab [] has joined #xen
06:37--- Netsplit <-> quits: tab, Beaky, plars, knewt, joink, drbyte
06:37--- Netsplit <-> quits: DEac-, monrad, athomas, JViz`, wenchien, matta-lt
06:37--- Netsplit <-> quits: Robot101, jonmason, demon, mael, Shaun, muli, surriel, schweeb, dst, apw, (+9 more, use /NETSPLIT to show all of them)
06:38--- Netsplit over, joins: schweeb, Robot101, muli, sunny, aliguori, apw, surriel, demon
06:38--- Netsplit over, joins: jonmason
06:38--- Netsplit over, joins: wenchien, monrad, Shaun, Sir_Ahzz, plars, Arnald, joink, JViz`
06:38--- Netsplit over, joins: cf0989b8, Bluefox, dst, DEac-
06:38--- Netsplit over, joins: buggs, mael, knewt, matta-lt, athomas, Beaky, drbyte, tab
06:38--- Channel: ServerMode/#xen [+o Sir_Ahzz] by
06:38--- Netsplit over, joins: caker, Tv
07:05--- ---> knewt_ [] has joined #xen
07:08--- <<-- knewt_ [] has quit (Quit: )
07:08--- ---> knewt_ [] has joined #xen
07:10--- <<-- knewt [] has quit (Killed (NickServ command used by knewt_))
07:10--- User: *** knewt_ is now known as knewt
07:12--- User: *** knewt is now known as knewt_
07:13--- User: *** knewt_ is now known as knewt
08:26aliguori ll
08:26aliguori asfd
08:31--- <<-- aliguori [] has quit (Read error: Connection reset by peer)
08:46--- ---> rharper [] has joined #xen
09:05--- <<-- monrad [] has quit (Quit: Leaving)
09:29--- ---> rtc2 [] has joined #xen
09:30rtc2 Hello! Is is possible to use xen on a system with linux 2.2 and egcs compiler to run linux 2.6?
09:33demon er, you need to have a 2.4.x or 2.6.x kernel patched with the appropriate patches to support the Xen virtual architecture
09:33rtc2 does that mean I need linux 2.4 or linux 2.6 to *run* xen?
09:33--- ---> jeroney [] has joined #xen
09:33mael yes
09:34demon well, you need to have a linux 2.4 or 2.6 kernel to run as domain 0 to control the rest
09:34demon so yeah
09:34demon unless support has been added to NetBSD 2 to act as dom0
09:34mael demon: you also need a paravirtualisation aware kernel for the domU so 2.2 is not an option
09:35mael your unprivileged domains must also be "ported" kernels
09:35demon right, as I said, you need to have an appropriate kernel patched with the relevant patches and built for the Xen hypervisor arch...
09:35rtc2 So xen does not run exclusively as a user process?
09:35rtc2 Sorry to ask, the web site is not very clear about it.
09:35demon Xen is a hyperkernel, it runs beneath all hosted domains
09:35demon between the hardware and the OS(es)
09:35mael xend is in userspace but not the hypervisor
09:36demon xend just provides the tools to control xen itself
09:36mael yeah exactly
09:37rtc2 The website could say so a little bit more explicitely.
09:37mael mmh it seemed clear to me
09:37mael though a picture or two could have been welcome :)
09:38mael but I must admit I read "xen and the art of virtualisation" also
09:38rtc2 From what does it follow xen is not exclusively a program for running modified kernels in userspace on the web site in your opinion?
09:39mael well you system boots on xen.gz which is the hypervisor itself
09:39mael then it boots the privileged domain dom0
09:40demon it runs modified kernels, but not in userspace
09:40rtc2 I mean on the main page,
09:41rtc2 You should definitively clearify the difference to things like User Mode Linux I think, as many people probably get confused
09:41--- ---> homebaum [] has joined #xen
09:42rtc2 BTW, how does xen compare to L4linux running on L4? Can it be compared?
09:42mael well I'm just a xen user, but I guess we can ask xen's devs to add a clear statement on what is xen :)
09:42mael I dunno L4linux
09:43rtc2 L4linux is a linux patched for running on L4, just as your patched linuxes are linuxes for running on xen...
09:43mael so I can't tell
09:43demon rtc2: did you read the FAQ page that's linked from the Xen main page?
09:43mael mmh but you don't have virtualisation with L4linux, do you?
09:44rtc2 mael: I don't know, never used it.
09:44* mael is quickly parsing l4linux web site
09:45mael mmh it seems you can have several linux kernel running on top of the L4 microkernel
09:46mael so from a high level point of view I guess you might consider that they have similar usage
09:48rtc2 demon: I did so.
09:48demon I guess the site isn't very clear on what L4Linux actually is/does
09:48rtc2 demon: Did I overlook something that answers my question?
09:49rtc2 demon: But I think the l4linux web page is quite clear about what l4linux is... "L4Linux is a port of the Linux kernel to the L4 microkernel. Ports for Linux versions 2.6, 2.4 and 2.2 are available."
09:49rtc2 it is the very first sentence on
09:49demon I've never heard of the L4 microkernel
09:50rtc2 is the concept of a microkernel known to you?
09:50demon yes, I was educated on it as part of my CS degree
09:51rtc2 Then it's sad you never heard of L4, because it's said to be the most promising microkernel.
09:51* mael heard of the hurd 3rd generation kernel being a L4 but that's all
09:51--- ---> soffi [] has joined #xen
09:51mael hey soffi
09:52demon well, other than CMU Mach (one of the better known microkernels), I'm not really familiar with what's out there... mostly because microkernels aren't that interesting to me
09:52mael Mister THE expert in ML's flaming :)
09:52soffi hey mael
09:52mael did you get my mail?
09:52soffi Lemme check
09:52rtc2 even hurd is now available for L4.
09:53soffi mael: I haven't checked my gmail since late monday... been stuck in a snowstorm and just got here late yesterday
09:53mael ok
09:54soffi ahh there it is
09:54mael a sarge domU image conformant to the draft specs is ready :)
09:54soffi allright
09:54--- ---> hollis [] has joined #xen
09:55mael I used it as a "concept validation"
09:55mael rtc2: well you just gave me a full load of webpages to read :)
09:55soffi ok you wanna scp them ?
09:55mael yeah I might
09:55mael same stuff?
09:56soffi yup
09:57soffi mael: I just read the netbsd howto.. and according to that you can boot a NetBSD cdrom image in XenU
09:57mael ok I transfer it then
09:57mael that's cool
09:58mael ok UL just started
09:58--- ---> deac [] has joined #xen
09:58mael soffi: this is something like the cow images on xen live cd?
09:59soffi mael: the netbsd stuff ?
09:59mael yeah
09:59mael the boot-on-cd stuff more precisely :)
09:59soffi have no idea, just glazed at it before going to bed lastnight
10:00mael I think I'll have a look at it then
10:01soffi If I understand it correctly you can create a domain with two(or more) block devices.. the first one being a cd-rom image for netbsd and the launch the domainU and get a NetBSD installer
10:01soffi then go through that... you have netbsd installed on the second vbd and then change the domU config file to remove the first vbd (which was the cd-img)
10:02mael hmm ok
10:02soffi or maybe I have been smoking crack
10:02mael :)
10:02--- <<-- DEac- [] has quit (Ping timeout: 480 seconds)
10:02mael I've read the cow stuff diagonaly and I think it is something approaching
10:02rtc2 Anyone been using User Mode Linux?
10:03demon we used to for some internal development systems
10:03mael a mean to boot on a read-only image and have the changes copyed on a read-write partition
10:03demon but we switched to xen
10:03rtc2 demon: How does it compare to xen?
10:04demon well, basically you're suggested to not allow loadable modules... I/O is a lot slower because it's pure userspace
10:04demon there are assorted caveats to its use
10:05demon it'll run standard linux/i386 binaries, but the performance is a killer
10:05rtc2 so its really only useful for kernel developement?
10:05rtc2 not a good idea to use for virtualization?
10:06mael not for production from what I heard
10:06demon no, because it's relatively easy for anything that might get root privilege inside the virtual environment to break out into the host environment
10:06rtc2 Okay, thanks for answering my questions.
10:06rtc2 Cya!
10:07demon whereas under xen, the domains are pretty well isolated
10:07--- <<-- rtc2 [] has quit (Quit: Leaving)
10:07mael mmh
10:08mael do you recon we should rename this chan to "#xen-drivein"?
10:08soffi hehe
10:09soffi I wanna be op on #xen-marketing
10:09mael lol
10:09mael no no xen-marketing is for the ML dude
10:09soffi ohhh
10:09soffi so I don't get to be op?
10:10mael no it's just that irc is too difficult for marketing ppl :)
10:10soffi hehe
10:10mael no i'm wrong
10:11mael it's just not hype enough
10:11mael they need IM
10:11mael with full color smileys
10:11mael they *have* to create an account to give their mail address to spammers' databases
10:12mael otherwise it's useless
10:12soffi hehe
10:14* mael sometimes wonder why nobody has yet started to enhance the irc protocol to had video and voice stuff
10:14soffi that's what iChat AV is for :)
10:14mael maybe everyone is just as lazy as I am :)
10:14* soffi own a powerbook
10:15soffi ghehe
10:15mael so do I
10:15mael but I'll never use ichat
10:15mael I use Xchat :)
10:15soffi hehe me to i use the XChat Aqua whateva
10:16mael yep
10:16* soffi is writing the ugliest php script ever
10:50--- ---> dtrumb [] has joined #xen
10:50buggs ...Failed to obtain physical IRQ 203
10:50buggs anyone ?
10:51@Sir_Ahzz hmmm?
10:51@Sir_Ahzz dom0 kernel handles that part.
10:51@Sir_Ahzz what's the drive that wants that irq?
10:51buggs with unstable when configuring Network interfaces
10:51--- ---> niv [] has joined #xen
10:52@Sir_Ahzz any network card? or a specific one, or a specific slot?
10:52buggs the builtin tigon3
10:52buggs it's a notebook
10:52* knewt rofl #
10:53@Sir_Ahzz does native linux find and use the network card just fine?
10:53buggs yes
10:53@Sir_Ahzz same kernel version as the xenlinux kernel.
10:53buggs yes
10:53--- <<-- Beaky [] has quit (Quit: no reason)
10:54@Sir_Ahzz what's the specific error messages the kernel spits out about that driver when it loads it? (or is it compiled in directly?)
10:54buggs compiled in
10:55buggs in testing and 2.0 from april 1st i can bring the the0 up but not actually do anything not even arp
10:55buggs *eth0
10:55@Sir_Ahzz odd.
10:55buggs anything i caan do to debug ?
10:56--- <<-- hollis [] has quit (Quit: leaving)
10:56@Sir_Ahzz not certain.
10:57@Sir_Ahzz emailed the xen-users mailing list with a copy of your kernel dmesg output?
10:57--- ---> cbrown [] has joined #xen
10:57buggs not yet, will try
11:13--- <--- niv [] has left #xen ()
11:15--- ---> hollis [] has joined #xen
11:45--- ---> monrad [] has joined #xen
11:49--- <<-- cbrown [] has quit (Quit: Leaving)
12:08--- <<-- athomas [] has quit (Quit: Leaving)
12:15--- <<-- soffi [] has quit (Ping timeout: 480 seconds)
12:22--- ---> philipp_ [] has joined #xen
12:34--- ---> Method [] has joined #xen
12:38--- <<-- sunny [] has quit (Quit: Changing server)
12:40--- ---> sunny [] has joined #xen
13:07--- <<-- dtrumb [] has quit (Quit: dtrumb has no reason)
13:26--- ---> niv [] has joined #xen
13:30--- <<-- matta-lt [~matta@] has quit (Ping timeout: 480 seconds)
13:57hollis I'm surprised people have put up with xend for this long. I haven't used it, but it sounds ridiculously fragile
13:58plars It is
14:02--- ---> lilo [] has joined #xen
14:52niv hence the anthony rewrite (vm tools)..
14:53@Sir_Ahzz xend was a fast-write type program. meant only for testing phases, but didn't get rewritten with 2.0. 8-P
14:54hollis fast-write?! and how many lines of code is it?
14:54@Sir_Ahzz it's in python.
14:55@Sir_Ahzz and not a lot of work was put into it.
14:55hollis "it's in python" means fewer lines of code
14:56niv xen needs to be *more* careful with memory, not less..
15:00@Sir_Ahzz agreed.
15:00@Sir_Ahzz hence the rewrite in C by soemone. :)
15:01@Sir_Ahzz they didn't want to have to worry about maintaining C code tools, prefered to use python for rapid development of the tools.
15:04niv true, there are tradeoffs, and it depends on the people and their skill set(s), which they are more comfortable with...
15:05niv i am having a wicked funny time with memory..forget about doing any network intensive test without at least 256 MB, pref 512..
15:05@Sir_Ahzz ?
15:06niv It does make a big difference to performance..
15:07niv I'll throw some results out to the list later..
15:08niv jon ran into this to some extent too, and I think I need to quantify it
15:09@Sir_Ahzz k
15:10Robot101 Sir_Ahzz: what was that userland iscsi daemon?
15:12@Sir_Ahzz ietd
15:13@Sir_Ahzz it's on sourceforge
15:13@Sir_Ahzz iscsi enterprise target daemon
15:22--- ---> tierra [] has joined #xen
15:46--- ---> anthony [] has joined #xen
15:46--- User: *** anthony is now known as aliguori
15:48--- <<-- deac [] has quit (Remote host closed the connection)
15:49plars niv: what network intensive tests are you running and having these memory issues?
15:51niv plars: netperf tcp stream msgsize - 8K (>)
15:52--- ---> DEac- [] has joined #xen
15:56--- <<-- DEac- [] has quit (Quit: )
16:07--- <<-- philipp_ [] has quit (Quit: C-x C-c)
16:12--- <<-- monrad [] has quit (Quit: Leaving)
16:27--- <<-- Method [] has quit (Read error: Operation timed out)
17:02--- ---> Method [] has joined #xen
17:26--- <<-- aliguori [] has quit (Quit: Leaving)
17:33--- <<-- jeroney [] has quit (Quit: Leaving)
17:52--- ---> monrad [] has joined #xen
18:02--- <<-- hollis [] has quit (Quit: Lost terminal)
18:02--- ---> cbrown [] has joined #xen
18:12--- ---> aliguori [] has joined #xen
18:19--- <<-- rharper [] has quit (Quit: Leaving)
18:27--- ---> eigood [] has joined #xen
18:55--- <<-- tierra [] has quit (Quit: Everybody wants to go to heaven, but nobody wants to die.)
19:12--- <<-- monrad [] has quit (Remote host closed the connection)
19:14--- ---> monrad [] has joined #xen
19:16--- <--- cbrown [] has left #xen (Leaving)
19:39--- <--- niv [] has left #xen (Leaving)
20:26--- <<-- Method [] has quit (Quit: .)
21:16--- ---> Nigelenki [] has joined #xen
21:16--- <<-- Bluefox [] has quit (Read error: Connection reset by peer)
21:38--- <<-- Nigelenki [] has quit (Ping timeout: 480 seconds)
21:39--- ---> hollis [] has joined #xen
21:41--- <<-- aliguori [] has quit (Quit: Leaving)
21:49--- ---> Nigelenki [] has joined #xen
22:12--- <<-- homebaum [] has quit (Quit: Client exiting)
23:36--- <<-- hollis [] has quit (Quit: leaving)
23:58--- ---> niv [] has joined #xen
23:59--- <--- niv [] has left #xen ()
---Logclosed Fri Apr 08 00:00:47 2005