[02:02] <Balsaq> youtube videos play slow on my old dell with buntu, was told old ATI video card do not have driver support for buntu, found an nvidia card that fit my system and nvidia says they have the buntu drivers, i am a non tech and a buntu newbie so i have not purchased the vid card yet-please comment
[02:05] <Balsaq> question:can i newer better vid card with the correct driver support for buntu, spped up online vidoe performance? or is all this controlled my the processor and ram and the vid card simply not involved in the speed of the vid?
[13:38] <cemoi> hello
[13:38] <cemoi> can i speak french?
[13:39] <cemoi> i search a chan fr for edubuntu
[13:54] <dgroos> Quick question: I did ifconfig and both nics say, "Bcast: and give an ip address range.  does that mean that both nics are broadcasting ip addresses?
[13:55] <dgroos> that would be a disaster...
[13:56] <ogra> indeed they broadcast ... every NIC does ... else your tcp/ip protocol wouldnt work
[13:56] <ogra> it has nothing to do with dhcp broadcasting
[13:56] <dgroos> (I can breath again)  Thanks ogra!
[13:57] <ogra> if you are 192.168.1.2 and want to ping 192.168.1.3 then your NIC first sends a request to the boradcast address to find out where 192.168.2.3 is
[13:57] <ogra> *err
[13:57] <ogra> 192.168.1.3 :)
[13:58] <ogra> if you use ping with the -b switch it will use the boradcast address to ping all available ip's in your network
[14:00] <dgroos> that makes sense though didn't know there were a/b switches--just googled.  That's good to know.
[14:02] <dgroos> ah... a/b is for cable/tv switches :)  /b /c /d /n /w /x /4 switches are for networks.
[14:04] <ogra> err, no i meant the commandline option -b for the ping command :)
[14:14] <dgroos> A perfect example showing that accurate communication isn't trivial!
[17:03] <LaserJock> is everybody testing Edubuntu Beta?
[17:11] <sbalneav> I kicked off the download last night
[17:12] <sbalneav> I'll probably try to install tonight.
[17:15] <LaserJock> great
[17:35] <alkisg> Hi sbalneav... do you think I have any chances in implementing the libpam-sshfs module that I told you the other day?
[17:35] <alkisg> Or will that be too hard, and I shouldn't even bother?
[17:37] <sbalneav> Well, it's certainly possible.
[17:37] <sbalneav> Here's the problem
[17:38] <sbalneav> Past PAM, you've also got NSS worries to contend with
[17:39] <alkisg> How are those handled? By seperate modules?
[17:39] <alkisg> I could reuse some existing libpam-something code for that, couldn't I?
[17:39] <sbalneav> Well.
[17:39] <sbalneav> (hmm)
[17:40] <sbalneav> Here's how *I* do it, but it has an inherent problem.
[17:43] <sbalneav> Create the pam module, collect the userid and password, and then initiate the ssh connection.
[17:43] <sbalneav> BUT have the ssh remain open, and have it create a control socket (the way we do with ldm in ltsp)
[17:43] <sbalneav> THEN
[17:44] <sbalneav> create an nss module that uses the established ssh session's control socket to look up passwd and group info from the server (via the getent command)
[17:44] <sbalneav> Problems:
[17:45] <sbalneav> 1) Password expiry.  It's what we face with LDM.  ssh doesn't handle expired passwords in a "graceful" manner.
[17:46] <sbalneav> i.e. there's no API that a pam module could tie into, you MUST actually parse the stdio stream with ssh, looking for the expired password messages.
[17:46] <sbalneav> problem is, this is screen scraping.
[17:46] <alkisg> OK, not a big problem for my needs
[17:46] <alkisg> e.g. the way ltsp handles it is good enough for me
[17:46] <sbalneav> 2) nss lookups are going to be very slow.
[17:47] <alkisg> About that. Do I really need to get all the passwd info from the server? E.g. does ldap also read the whole database, even if there are thousands of users in there?
[17:47] <sbalneav> However, you may be able to get around that by cacheing.
[17:47] <alkisg> Or could I get away with only providing that one user locally?
[17:48] <sbalneav> libnss_ldap does an ldap query for every call that would require info on "human readable" names for files
[17:48] <sbalneav> Well, here's what I'd do.
[17:48] <sbalneav> I'd simply have the pam module, upon successful login, to a "getent passwd", and cache the file locally
[17:49] <sbalneav> then just use libnss_extrafiles to go after that local cache.
[17:52] <alkisg> Thanks man... ok, I'll try this, I do hope it's doable in some reasonable time.... isn't it? :D
[17:58] <alkisg> Thanks again :)
[18:58] <dgroos> alkisg: I used the info you gave yesterday about switching the nics and it worked!
[18:58] <alkisg> dgroos: Nice!! Put it on the ubuntu wiki if you want
[18:59] <dgroos> I first had to fix some stuff I'd messed around with before doing it right :)
[18:59] <dgroos> Great idea--I'll do it.
[19:00] <dgroos> What name would you give the page? AssigningLabelsToRecalitrantNics?
[19:00] <dgroos> ;)
[19:02] <dgroos> And, do I do it at help.ubuntu.com/community  hmmm... any thoughts?
[19:53] <alkisg> https://help.ubuntu.com/community/UbuntuLTSP/ServerCloning ?
[19:54] <alkisg> (assuming that sometime someone will fill more info about it :D)