=== jibel_ is now known as jibel | ||
ogra_ | dm8tbr, so is there a chance you can get us the syslog ? | 10:21 |
---|---|---|
dm8tbr | ogra_ roped me in here... On current daily-live amd64 ubiquity dies before even displaying the live/install choser. known problem? | 10:21 |
dm8tbr | as soon as I get home, but I also created one bug with an image from a few days ago that contained a lot of stuff | 10:22 |
dm8tbr | let me see if I can find it | 10:22 |
ogra_ | ah, yeah, that would help :) | 10:22 |
dm8tbr | narf don't have my launchpad credentials on this laptop. just a sec | 10:27 |
dm8tbr | 1153272 1153266 | 10:27 |
dm8tbr | the other one is for network-manager doing very weird stuff and crashing | 10:27 |
ogra_ | bug 1153272 | 10:27 |
dm8tbr | I'll be able to try things again once I come home | 10:27 |
ogra_ | bug 1153266 | 10:27 |
ogra_ | come on bot ... | 10:27 |
dm8tbr | :)2~ | 10:27 |
dm8tbr | slow bot is slow | 10:28 |
ogra_ | heh, yeah | 10:28 |
ogra_ | or dead | 10:28 |
dm8tbr | or lobotomized | 10:28 |
ogra_ | heh | 10:28 |
dm8tbr | (that's a real setting for supybot, which ubottu is based on. it mutes the bot on a given channel) | 10:28 |
ogra_ | well, i doubt it was actuvely muted | 10:29 |
ogra_ | *actively | 10:29 |
dm8tbr | given that the hardware is not totally super ancient (ThinkPad T61p), I guess this should be of interest even if it's somehow hw-related | 10:30 |
ogra_ | yeah | 10:30 |
dm8tbr | oh and just in case: it happened with 3 different daily images and two different usb-sticks | 10:31 |
ogra_ | looks like the first one is NM related as well | 10:32 |
dm8tbr | interesting, so nm crashes ubiquity | 10:33 |
* dm8tbr puts bets on the intel wifi card | 10:34 | |
dm8tbr | iw3945 IIRC in that model | 10:34 |
ogra_ | Mar 10 14:49:51 ubuntu ubiquity[3145]: File "/usr/lib/ubiquity/ubiquity/nm.py", line 210, in build_cache | 10:35 |
ogra_ | Mar 10 14:49:51 ubuntu ubiquity[3145]: devices = self.manager.GetDevices() | 10:35 |
ogra_ | thats in the traceback of the first one | 10:35 |
ogra_ | so it dies when talking to NM it seems ... falling over in dbus communication | 10:36 |
ogra_ | if you have a HW switch to disable the wifi i would try another install with it disabled ... if it works you can be pretty sure its the wlan card | 10:37 |
ogra_ | the second bug points to eth0 though | 10:37 |
ogra_ | Mar 10 14:34:13 ubuntu NetworkManager[9700]: <info> (eth0): deactivating device (reason 'managed') [2] | 10:38 |
ogra_ | Mar 10 14:34:13 ubuntu NetworkManager[9700]: <error> [1362926053.564506] [nm-system.c:254] sync_addresses(): (eth0): error -19 returned from rtnl_addr_delete(): Invalid address for specified address family | 10:38 |
dm8tbr | interesting enough, killing NM and running dhclient works just fine | 10:40 |
dm8tbr | I'll have a look though | 10:40 |
ogra_ | yeah, and a recent log woould be good | 10:41 |
ogra_ | to see if its still the same issue | 10:41 |
dm8tbr | should we discuss this still here or does the NM component make it a different problem? | 10:42 |
ogra_ | well, it tears down ubiquity ... | 10:51 |
dm8tbr | bug 1153272 | 13:07 |
ubot2 | Launchpad bug 1153272 in ubiquity (Ubuntu) "ubiquity crashed with dbus.exceptions.DBusException in call_blocking(): org.freedesktop.DBus.Error.ServiceUnknown: The name :1.37 was not provided by any .service files" [Medium,New] https://launchpad.net/bugs/1153272 | 13:07 |
dm8tbr | ogra: ah, now it works. The bug was marked private, because the information collector thought it had picked up password data in the logs | 13:08 |
ogra_ | ah | 13:08 |
xnox | dm8tbr: that's a dupe of our never-ending whoopsie/nm/ConsoleKit timing out. | 13:08 |
ogra_ | still ? | 13:08 |
ogra_ | i thought that was fixed weeks ago | 13:09 |
xnox | but it shouldn't be happening on the (20130310) image (from that report) | 13:09 |
ogra_ | right | 13:09 |
dm8tbr | xnox: It happened with todays image too | 13:09 |
dm8tbr | nm goes into an endles loop of connect/disconnect | 13:10 |
dm8tbr | bug 1153266 is now public too | 13:10 |
ubot2 | Launchpad bug 1153266 in network-manager-applet (Ubuntu) "nm-applet crashed with SIGSEGV in add_to_object_array_unique()" [Medium,New] https://launchpad.net/bugs/1153266 | 13:10 |
dm8tbr | haha, I think this is an ipv6 problem | 13:14 |
dm8tbr | nm_setting_ip6_config_get_dhcp_hostname: assertion NM_IS_SETTING_IP6_CONFIG (setting) failed | 13:15 |
dm8tbr | then two more messages and then the kernel says segfault | 13:15 |
dm8tbr | I can take this to the NM people, I hang out on their channel anyway. Thoughts? | 13:16 |
xnox | ogra_: reading nm.py, we should be more resilient and expect nm to fail. but that shouldn't be bringing the installer down. Plus I have bugs to fix in nm.py to match recent design. | 13:16 |
xnox | cyphermox: ^ bug 1153266 | 13:16 |
ubot2 | Launchpad bug 1153266 in network-manager-applet (Ubuntu) "nm-applet crashed with SIGSEGV in add_to_object_array_unique()" [High,Confirmed] https://launchpad.net/bugs/1153266 | 13:16 |
xnox | dm8tbr: we have nm people hanging out here as well ;-) | 13:16 |
dm8tbr | ah, cyphermox is also here :) | 13:16 |
dm8tbr | just noticed :) | 13:16 |
cyphermox | yeah | 13:16 |
dm8tbr | let me see if I can find more info than the bug has | 13:17 |
dm8tbr | Mar 10 14:36:12 in https://launchpadlibrarian.net/133563713/WifiSyslog.txt | 13:18 |
dm8tbr | that's what I see now too | 13:18 |
dm8tbr | seems it tanks on trying DHCPv6 | 13:18 |
dm8tbr | my router _should_ only send RA, but who knows what Netgear did to it... | 13:19 |
dm8tbr | that also explains it why I'm the first to experience the bug and report it... mass v6 deployment and all that | 13:19 |
cyphermox | seems unlikely to be dhcpv6 | 13:20 |
cyphermox | could be the new DUID code though | 13:20 |
cyphermox | (I test ipv6 everyday here) | 13:20 |
cyphermox | http://paste.ubuntu.com/5613583/ | 13:21 |
cyphermox | perhaps I should change the lease and RA timeouts :) | 13:21 |
cyphermox | anyway, I'm going to test something, see if it's DUID | 13:21 |
xnox | In ubiquity: "# TODO: DBus exceptions. Catch 'em all." | 13:31 |
dm8tbr | cyphermox: let me dump some RA traffic | 13:32 |
dm8tbr | attached dump to the ticket | 13:45 |
dm8tbr | https://bugs.launchpad.net/ubuntu/+source/network-manager-applet/+bug/1153266/+attachment/3573991/+files/radvdump.txt | 13:45 |
ubot2 | Launchpad bug 1153266 in network-manager-applet "nm-applet crashed with SIGSEGV in add_to_object_array_unique()" [High,Confirmed] | 13:45 |
cyphermox | dm8tbr: I wonder if it has to do with your particular setup w/ other-flag, so the RAs for addresses and DHCP for dns and stuffs? | 14:27 |
cyphermox | stgraber: you have the rig to test ipv6, would you have time to reproduce this crash? ^^ | 14:27 |
dm8tbr | if you tell me what to do I can also do something | 14:28 |
cyphermox | well, we have some automated tests for ipv6 -- we shoud first check if it's easy to reproduce with that | 14:28 |
dm8tbr | I could e.g. dump all the ethernet traffic that goes to the system around crash time | 14:28 |
dm8tbr | ok | 14:28 |
cyphermox | nah, I don't think it will be necessary | 14:29 |
cyphermox | we do have a stack trace | 14:29 |
cyphermox | andthe three last messages in syslog before the crash | 14:30 |
cyphermox | dm8tbr: I filed the bug upstream, updated LP | 14:54 |
dm8tbr | k, thanks | 15:06 |
* dm8tbr feels useful for finding a bug in NM :) | 15:06 | |
ogra_ | yay | 15:12 |
=== kentb is now known as kentb-out | ||
buhman | stgraber: I noticed the same problem | 22:34 |
buhman | (http://irclogs.ubuntu.com/2013/02/27/%23ubuntu-installer.txt 20:25) | 22:34 |
buhman | stgraber: I managed to make it boot *once*, and this happened to be the time I was recording the boot sequence http://youtu.be/Qttov7aTzkA | 22:35 |
buhman | I've not since been able to make that happen | 22:35 |
buhman | normally I just get a single "error: couldn't send network packet." and that's it | 22:37 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!