=== asac_ is now known as asac [08:23] hi [08:33] http://www.phoronix.com/scan.php?page=article&item=ubuntu_bench_2008&num=1 [08:36] yeah ... that was posted somewhere ;) [08:43] sqlite is quite3 a big issue for us i think [08:47] some of those numbers seem flat-out silly to me [08:47] but i've long had issues w/ larabel's testing methodology [08:49] maybe [08:51] comparing sqlite 3.6.3 fr'example... great, but 7.04 had sqlite 3.3.13 [08:51] you can't say "feisty would be zomgfast if you just updated all the apps to the versions in intrepid" [09:04] compiler hardening and glibc checking are probably part of it too [09:09] ack [09:09] well ... sqlite getting slower is a different thing [09:10] sqlite for us (moz) or for the whole distro ? [09:10] its somehow not comprehensible how sqlite would require so much more CPU for the same [09:10] fta: for us and the distro [09:10] fta: it probably has the most severe impact for us though [09:11] as ffox uses it in a heavy way [09:11] it's getting better in 3.1 [09:11] yeah ... but still if sqlite would be twice as fast [09:12] things would become even more snappy [09:14] interesting thing is that rhythmbox always opens on _some_ desktop -- which is rarely the current active desktop [09:17] everything using the tray is broken for me, file copy in nautilus (gvfs), liferea, rhythmbox, deluge, ... except xchat [09:17] in which way? [09:18] like what i said for phythmbox? [09:18] fta: ? [09:18] once open in a workspace, it's impossible to rise it from another workspace [09:18] yeah [09:18] thtas true [09:18] fta: ++ [09:18] ;) [09:19] i think gtk is broken [09:19] fta: when did this start for you? [09:19] raise [09:19] a few months ago [09:19] did this happen in this last round of tarballs? [09:19] hmm [09:19] fta: or maybe its compiz [09:20] i fixed liferea with Jazzva once, it's broken again [09:20] or aint you using that? [09:20] i use metacity [09:20] ok so not compiz [09:20] i still bet its gtk ;) [09:21] fta: what did that liferea fix consist toof? [09:21] did someone drop it or is it just not applicable anymore? [09:24] http://paste.ubuntu.com/63562/ [09:25] looks like a hack [09:26] like something that is predestined to break ;) [09:26] yes, but GDK_WINDOW_STATE_ICONIFIED means different things [09:26] its dumb [09:26] yeah i can imagine that [09:26] i checked the doc at that time [09:27] if you feel like fixing all this... [09:27] i wonder why noone else complained [09:27] me? ... gtk folks think they are perfect ;) [09:28] fta: most likely there are a bunch of bugs open [09:29] fta: just not filed against gtk+2.0, but agains the individual apps [09:29] ok ... have to move ... bbi10min [13:16] asac: fta: plenty of room for mozilla team sessions for openweek: https://wiki.ubuntu.com/UbuntuOpenWeek [13:28] bummer === jdstrand_ is now known as jdstrand [16:15] [reed]: you think you could lend me an hour or so for getting info on the EAP thing? [16:16] <[reed]> sure [16:16] <[reed]> now is ok [16:17] <[reed]> asac: now ok for you? [16:23] [reed]: \o/ ;) yeah. [16:24] [reed]: so its still failing right? [16:24] [reed]: when trying to connect and it fails, what do you see in /var/log/wpa_supplicant.log (maybe to a tail it) [16:25] <[reed]> yes, though I've gotten it to work twice randomly (but never was able to get it working immediately after that... very sporadic) [16:25] <[reed]> but that's twice in like 2-3 weeks [16:25] <[reed]> :( [16:25] urgh. scares me :/ [16:25] [reed]: anyway ;) [16:26] [reed]: there is a bunch of users that report it as working with wpa_supplicant, so i hope we can find a fix [16:27] <[reed]> let me restart to get a blank slate [16:29] yeah. though lets hope that thats on necessary ;) [16:30] <[reed]> oh, have my weekly call with my manager now, so give me about 5-10 min. [16:31] [reed]: sure ;) [16:34] <[reed]> asac: he hasn't called yet, so I'll post some logs [16:34] <[reed]> give me a sec [16:34] cool ;) [16:37] asac: http://reed.pastebin.mozilla.org/563417 [16:38] reed: ok so what kind of network is that? PEAP? [16:39] otp [16:44] <[reed]> ok, back [16:44] <[reed]> yeah, PEAP [16:45] [reed]: ok. so there should be in syslog a few lines where NM dumps the parameters it sends to wpa [16:45] [reed]: do you see those? [16:45] http://reed.pastebin.mozilla.org/563420 [16:46] msu1x is 802.1x... msstate.edu is just a normal open wifi net that requires me to VPN [16:46] reed: so how was that connection setup? just started NM and selected the AP? or did you already tweak stuff manually in connection editor= [16:46] ? [16:47] no manual tweaking [16:47] just started NM [16:47] I can delete the two APs [16:47] and start from scratch [16:47] if it would help [16:47] reed: could you try to create a connection and manually configure stuff? [16:47] pretty sure I already tried that, though [16:47] reed: tried that? [16:48] hmm [16:48] not sure [16:48] ok, give me a sec to test [16:48] reed: yeah. if you look at wpasupp log [16:48] reed: you see that NM tries to force peapver=1 and wpasupplicant complains about that [16:48] might be a red herring, but in supplicant code it just return NULL; if it sees that [16:49] so might contribute here [16:49] reed: i looked at NM applet and the only other option available is peap version 0 ... so trying that would be a first stab [16:49] tried that [16:50] reed: can you try and see if that error goes away in wpasupplicant.log (this would mean its a red herring ;)) [16:51] <[reed]> yeah, I don't see any error about peap v1 in wpasupplicant.log [16:51] <[reed]> when using 0 [16:51] <[reed]> but still doesn't work [16:51] <[reed]> :) [16:53] [reed]: you have a log with ver 0? or is its just the same but without that error? [16:53] same but without error [16:53] http://reed.pastebin.mozilla.org/563423 is syslog of disconnect from msstate.edu, attempt connect to msu1x, fail [16:54] this is the only thing in wpasupplicant.log for msstate.edu -> msu1x -> fail [16:54] CTRL-EVENT-DISCONNECTED - Disconnect event - remove keys [16:54] No network configuration found for the current AP [16:54] CTRL-EVENT-DISCONNECTED - Disconnect event - remove keys [16:54] CTRL-EVENT-SCAN-RESULTS [16:55] reed: i think we need a more verbose wpasupplicant.log, but first lets try if it really works without NM [16:55] in /etc/network/interfaces add something like: [16:55] iface eth1 inet dhcp wpa-conf /path/to/wpa_supplicant.conf [16:55] oops [16:55] again: [16:55] iface wlan0 inet dhcp [16:56] wpa-conf /path/to/wpa_supplicant.conf [16:56] and create a wpa_supplicant.conf at the configured path with: [16:56] reed: http://paste.ubuntu.com/63679/ [16:57] <[reed]> I removed the ca line [16:57] <[reed]> what about peaplabel? [16:57] not sure :/ [16:58] reed: keep it ... and if it doesnt work remove ;) [16:58] <[reed]> ok, now what? [16:58] reed: so ... if you have that stop NM sudo /etc/init.d/NetworkManager stop [16:58] and try to ifup your interface [16:58] sudo ifup wlan0 (given wlan0 is your interface.name) [16:59] here is another PEAP example with label=0 http://linux.die.net/man/5/wpa_supplicant.conf [17:00] <[reed]> it's just doing DHCPDISCOVER a ton [17:00] <[reed]> never getting an IP [17:01] reed: bug 290337 [17:01] Launchpad bug 290337 in network-manager "Fails to connect to wireless network (dup-of: 272185)" [High,Triaged] https://launchpad.net/bugs/290337 [17:01] Launchpad bug 272185 in network-manager-applet "[Intrepid] iwl3945 + iwl4965 -- network-manager will not connect to a WPA EAP (Enterprise) network (disassociating by local choice (reason=3) )" [High,Fix released] https://launchpad.net/bugs/272185 [17:01] that contains a conf someone said worked [17:01] <[reed]> looking [17:01] reed: but its supposed to do dhcp right? or do you have static stuff? [17:01] <[reed]> supposed to do DHCP [17:01] <[reed]> let me get rid of phase1 [17:02] yeah [17:02] try that [17:02] there are bunch in the messy master bug (the second) that say that it works manually (e.g. like what we try now) [17:03] <[reed]> hmm, nm-system-settings and wpa_supplicant are both running even though I've ifdown'd the iface [17:03] <[reed]> is that ok? [17:03] reed: oh right [17:04] reed: settings is ok [17:04] but wpasupplicant might fight over the interface with the manual one [17:04] so killall wpasupplicant [17:04] check that none is runnig and try ifup wlan0 [17:04] (first ifdown wlan0 if it refuses to do stuff) [17:05] <[reed]> how can I make sure of the SSID I'm connected to? [17:05] reed: iwconfig [17:05] should spit out the current state [17:06] yep [17:06] I'm on msu1x [17:06] so works [17:06] cool [17:06] yep [17:06] so, it's NM's fault ;) [17:06] reed: obviously ;) [17:06] reed__: but thats good ;) [17:06] reed__: can you insert exactly what you see in the paste? [17:07] see in syslog or what? [17:07] reed__: http://paste.ubuntu.com/63686/ [17:07] reed__: i think the keys should match (maybe its _ replaced by -) [17:07] ah [17:07] ok [17:07] disconnecting... will try now [17:12] <[reed]> just getting DHCPDISCOVER [17:12] <[reed]> over and over [17:12] <[reed]> so, I guess you want me to comment out a row by row [17:13] <[reed]> and figure out which one is breaking it? :) [17:13] <[reed]> or do you have some ideas [17:13] [reed]: so what do you have now? [17:13] [reed]: i assume you didnt have group/pairwise lines before? remove those two first [17:13] and proto too i guess [17:14] <[reed]> network={ [17:14] <[reed]> ssid="msu1x" [17:14] <[reed]> scan_ssid=1 [17:14] <[reed]> key_mgmt=WPA-EAP [17:14] <[reed]> proto="WPA RSN" [17:14] <[reed]> pairwise="TKIP CCMP" [17:14] <[reed]> group="WEP40 WEP104 TKIP CCMP" [17:15] <[reed]> fragment_size=1300 [17:15] <[reed]> eap=PEAP [17:15] <[reed]> identity="myusername" [17:15] <[reed]> password="mypassword" [17:15] <[reed]> phase1="peapver=0" [17:15] <[reed]> phase2="auth=MSCHAPV2" [17:15] <[reed]> } [17:15] <[reed]> that's what I have right now (had to type it since my other laptop isn't connected right now) [17:15] <[reed]> I'll remove group, pairwise, and proto now [17:16] yeah [17:17] <[reed]> oh, just noticed this on ifup [17:17] <[reed]> wpa_supplicant: /sbin/wpa_supplicant daemon failed to start [17:17] [reed]: you need to ifdown first? [17:17] <[reed]> I did [17:17] [reed]: is a wpasupplicant running? [17:17] <[reed]> no [17:17] maybe kill it to be sure ;) [17:17] hmm [17:17] <[reed]> ps x | grep -i wpa [17:17] <[reed]> returns nothing [17:17] <[reed]> let me try again [17:18] [reed]: hmm i think you dont need the " for proto and pairwise and group [17:18] might even be illegal [17:18] as its a list of strings rather than one string [17:18] <[reed]> well, getting that error after removing those, too [17:19] example config suggests things like: group=CCMP TKIP WEP104 WEP40 [17:19] <[reed]> oh, maybe I didn't save that [17:19] <[reed]> ok, didn't get the error that time [17:20] <[reed]> still have group, proto, and pairwise in file [17:20] <[reed]> it's doing DHCPDISCOVER [17:20] <[reed]> let's see if it gets an IP [17:20] hopefully not ;) [17:21] http://reed.pastebin.mozilla.org/563434 [17:21] it got an ip :( [17:22] reed__: ok but took some time :/ [17:22] yeah [17:22] reed__: does it always work? [17:22] checking [17:22] like ifdown wlan0; ifup wlan0 ? [17:23] reed__: also how long does it about take to get that IP? [17:25] <[reed]> it worked again the second time, but not third time... trying again [17:26] <[reed]> didn't work that time either [17:26] reed__: checked that wpasupplicant isnt running ;) [17:26] ? [17:26] <[reed]> not running [17:26] reed__: ok. i think thats enough empirical data. when it worked, how long did it take? [17:27] <[reed]> took a while... at least 30 seconds... probably closer to a minute [17:27] more like 3 seconds or 10 or 30? [17:27] reed__: ok [17:27] reed__: can you drop the stuff we discussed now? [17:27] <[reed]> if it fails, it takes 1m14s [17:27] <[reed]> drop proto, group, pairwise? [17:27] <[reed]> and try again? [17:27] yes [17:27] <[reed]> ok [17:28] reed__: do you remember how long it took for the first attempt that worked? (e.g. the simple config)? [17:28] <[reed]> it was super quick [17:28] <[reed]> I remember that [17:28] cool [17:29] reed__: ok, lets drop the default group... stuff for now [17:29] <[reed]> ok, got an ip, but it took five DHCPDISCOVERs and `time ifup wlan0` returned 46.434s [17:30] [reed]: hmm. so thats with the dropped proto... stuff? [17:30] <[reed]> yes [17:30] <[reed]> that's with proto, pairwise, and group removed [17:30] reed__: do you still have the "simplish" config and could check whether its still snappy? [17:30] <[reed]> sure [17:30] maybe try two or three times [17:30] sometimes the driver is in bad shape ;) [17:32] <[reed]> hmm, didn't work [17:32] <[reed]> let me try again [17:37] <[reed]> can't get it to work again [17:37] <[reed]> :( [17:39] reed__: urgh .... so maybe driver state busted? [17:40] reed__: try ifdown wlan0 .... [17:40] sudo modprobe -v -r iwlang [17:40] and [17:40] sudo modprobe -v iwlang [17:40] <[reed]> iwlang not found [17:40] err iwlagn ;) [17:40] * asac remembers iwl-agn :) [17:41] <[reed]> worked... took 37.438s and four DHCPDISCOVERs [17:42] <[reed]> with simple config [17:43] reed__: ok. not really a sprint, but ok [17:43] reed__: can you try to do the same one more time for the full (with all) and the group/pairwise/.. stripped conf? [17:43] e.g. remove driver first [17:43] <[reed]> sure [17:48] <[reed]> I've apparently locked up my laptop [17:48] <[reed]> awesomne [17:49] <[reed]> yep [17:49] <[reed]> locked up [17:49] <[reed]> restarting now [17:49] <[reed]> grrr [17:49] <[reed]> "System policy prevents restarting the system when other users are logged in" [17:49] <[reed]> no, that's such a lie [17:49] <[reed]> I'm the only one logged in [17:50] yeah sorry. doing driver stuff always has the risk [17:50] of locking stuff up [17:50] intel drivers are just getting worse the more those folks work on them [18:00] <[reed]> asac: so, group, pairwise, etc. removed and driver restarted took 44.393s [18:00] <[reed]> and worked [18:03] [reed]: doesnt sound convincing ;) [18:03] <[reed]> yeah [18:04] [reed]: ok thanks for testing now. most likely its a combination of driver/timeout tuning/dont passing defaults for group, ... when not needed [18:04] will bring up a test package for the latter two soon and let you know ;) [18:06] <[reed]> ok [18:11] [reed]: ok so the timeout is 25 seconds. if you want you could spin your own network-manager package and see if tuning that to ugly 60 seconds helps you to connect more frequently ;) [18:11] the other thing i will try is to remove the defaults for group/pairwise, etc. so in the end its quicker [18:13] asac: Trying to set the default homepage for firefox .. what is the best way .. I hear you would know [18:13] [reed]: http://paste.ubuntu.com/63698/ [18:13] [reed]: apt-get source network-manager; sudo apt-get build-dep network-manager; cd network-manager-*/; patch -p0 < /tmp/thispatchthings.txt; dpkg-buildpackage -rfakeroot -b [18:14] asac: we are trying to set the homepage to be "www.igoogle.com" (which is mozilla blessed) .. but there is no clear way to do it .. tyring /etc/firefox-3.0/prefs/firefox.js turns up no luck .. ubufox seems to override setting the homepage [18:14] <[reed]> asac: ok, trying [18:14] jerone-mobile: yeah. in intrepid there is an example in /etc/firefox-3.0/pref/ubufox.js [18:15] jerone-mobile: also look at the .properties fiel reffed therein [18:16] [reed]: i hope i catched the right timeout-knob ... there are a bunch of different timeouts, but from code this looks like it and the syslog you pasted also seems to be 25s when it gives up [18:17] asac: ok tried that .. this is with hardy .. and it didn't seem to work I put: pref("browser.startup.homepage","http://www.igoogle.com"); in /etc/firefox-3.0/pref/firefox.js [18:17] jerone-mobile: then look closer ;) [18:17] the example is different [18:18] asac: ah I see has to be a properties file [18:18] jerone-mobile: yeah [18:19] [reed]: -p0 is of course none-sense ;) -p1 but i guess you found that out ;) [18:24] asac: one more slot left for openweek, sure you don't want it? [18:24] asac: you can do 3g! [18:27] <[reed]> asac: gimme a bit... I'm freezing [18:27] * [reed] gets a coat [18:27] <[reed]> it's 50F [18:27] <[reed]> :( [18:28] jcastro: thanks for the reminders. i think i will skip this open week (i participated every week so far!!). too busy and i will start a holiday in that week ;) [18:28] yeah, I hear ya. [18:29] next time i will swamp you ;) [18:30] heh [18:32] <[reed]> asac: it's building now [18:34] <[reed]> asac: ok, does dpkg-buildpackage install it and such? [18:36] [reed]: no [18:36] [reed]: cd ..; sudo dpkg -i network-manager*.deb libnm-*.deb [18:37] <[reed]> ok [18:37] [reed]: then comment stuff in /etc/network/interfaces (otherwise your device would be unmanaged in NM) and restart NM [18:37] <[reed]> ok [18:37] sudo /etc/init.d/NetworkManager restart ;) [18:37] <[reed]> yeah [18:38] <[reed]> I know that :p [18:38] [reed]: if it doesnt work oob give driveres a chance and restart system [18:38] asac: hmm.. well even doing the file way it seems ubufox takes over the homepage and makes it start.ubuntu.com/8.04 .. maybe this is something that only works with the Interpid build .. I'm trying to get it working in UNR 1.01 (based on hardy) with firefox-3.0.3 .. but no love .. any other ways? [18:38] [reed]: you can verify that the new timeout is used by checking that the applet spins for about a minute obviously [18:39] jerone-mobile: that "ubufox" win thing is fixed in intrepid, yes. [18:39] jerone-mobile: you can backport that fix to your ubufox package easily though [18:39] asac: ok good to know I can ask someone to backport it [18:40] jerone-mobile: give the guy who does this the following info: cherry-pick rev 80-81 from branch https://code.edge.launchpad.net/~ubuntu-core-dev/ubufox/ubuntu.intrepid ... to your packaging branch [18:41] asac: big thanks! [18:41] welcome [18:41] jerone-mobile: the ubufox win bug is 259914 [18:41] (as documented in changlog/commit) [18:43] <[reed]> asac: wooo [18:43] <[reed]> it connected [18:47] [reed]: how long did it take? [18:49] <[reed]> a whiler [18:49] [reed]: also please double check that this still works after reboot and i will send out packages to the bug-mob ;) [18:49] <[reed]> yeah [18:49] <[reed]> I will [18:49] <[reed]> going to class [18:49] <[reed]> will try there [18:49] [reed]: its ridiculous ... 45seconds to connect [18:50] [reed]: cool [18:50] thanks a bunch [18:50] too late for final unfortunately, but i am currently accumulating a bunch of fixes for a stable update to send out in the beginning [18:51] asac: are you and reed talking about wireless time out .. I see this issue actually from time to time .. it's not giving it enough time to get the dhcp address .. but then I try it a second time and it works [18:51] jerone-mobile: what kind of net? [18:51] asac: wireless [18:51] asac: wireless 802.11n with WPA2 [18:51] jerone-mobile: lol [18:51] jerone-mobile: what kind of wpa? [18:51] eap? [18:51] or psk? [18:51] which driver? [18:52] asac: iwlagn [18:52] yeah ok [18:52] jerone-mobile: is that PSK? [18:52] asac: eap [18:52] jerone-mobile: what kind of eap? [18:52] peap? ttls? [18:53] jerone-mobile: do you run that in the office? (are you in an office at all?) [18:53] asac: actually it's just WPA2 Personal [18:53] jerone-mobile: ok [18:53] jerone-mobile: thats even more crazy [18:53] jerone-mobile: multiple APs with the same ESSID? [18:53] e.g. roaming? [18:53] asac: nope just one [18:54] jerone-mobile: whats you lp id? [18:54] asac: I've seen it 3 times .. and it's allways after a cold boot. [18:54] asac: jerone [18:54] jerone-mobile: ok i subscribe you to a WPA-EAP bug where i will send timeout tuned packages soon ;) [18:54] done [18:54] bug 272185 [18:54] Launchpad bug 272185 in network-manager-applet "[Intrepid] iwl3945 + iwl4965 -- network-manager will not connect to a WPA EAP (Enterprise) network (disassociating by local choice (reason=3) )" [High,Fix released] https://launchpad.net/bugs/272185 [18:54] asac: ok [18:55] jerone-mobile: not sure. but if its really timeout it should also fix you [19:04] [reed]: another interesting thing to see would be if things become much quicker when only the right values for group/pairwise/... are setup in wpasupplicant.conf [19:04] [reed]: if you feel like playing around with it, I would appreciate that input [19:05] (reason: when its much quicker then its a wpasupplicant bug, otherwise most likely a driver thing) === fabrice_sp_ is now known as fabrice_sp [20:47] asac: so it look like customization is actually broken in intrepid .. while ubufox no longer over rides the homepage .. if you try to set the homepage .. you get no default home page at all [20:50] jerone-mobile: well. it works here ;) [20:50] jerone-mobile: just remove the comment from the example [20:50] asac: hmmm [20:50] that should work [20:50] the example .properties is also there [20:51] asac: yeah I'm trying it on my laptop now and when I add an entry into ubufox.js .. I get nothing .. yeah I did the example file way .. let me do an update [20:55] an update? [20:55] jerone-mobile: http://paste.ubuntu.com/63758/ [20:55] thats what i have [20:55] jerone-mobile: take care that you must not have a "local" ubufox installed in your profile [21:09] asac: ok I got it .. not sure what I was doing wrong .. but thanks [21:18] jerone-mobile: cool. welcome. [21:19] W: GPG error: http://ddebs.ubuntu.com intrepid Release: The following signatures couldn't be verified because the public key is not available: NO_PUBKEY ECDCAD72428D7C01 [21:19] \o/ [21:26] fta: yeah. final bump [21:26] fta: paste that in -devel ;) ... maybe someone reacts [21:26] now, same thing for ppas would be nice [21:26] fta: huh? [21:27] ddebs repo was not signed [21:27] ah ;) [21:27] ok [21:27] * asac never noticed (scary) [21:27] oh yeah i noticed ;) [21:27] now that you say it ;) [21:28] ok ... using NM on a server install ;) [21:28] now i just have to think whether i want to migrate my mailsystem to that thing [21:30] i upgraded my desktop at work from hardy to intrepid, now, it's dead [21:31] fta: yeah. you should have used NM ;) [21:31] it was fine up to the reboot, but now, i can't login, neither gdm nor console. [21:31] i mean 0.7 :) [21:32] nope, it's not the network [21:32] fta: you say it also is dead when sitting in front of it? [21:32] it seems there's not enough memory to start the login shell [21:32] sounds ... erm, unlikely [21:32] indeed [21:32] fta: what happens? [21:33] it's stuck after the password [21:34] i will retry tomorrow, but now, the box ping but i can't ssh it [21:34] fta: use nfs? [21:34] nope [21:34] for home or something?` [21:34] ok [21:34] then no clue [21:34] me neither [21:34] when you see that you should report it i guess [21:35] it's difficult to explain [21:47] huge split, we were just 3 on my side :) [21:54] http://ppa-search.appspot.com/ [22:06] @time [22:06] Current time in Etc/UTC: October 28 2008, 22:06:40 - Next meeting: QA Team in 18 hours 53 minutes === ubott2 is now known as ubottu