[08:53] guys, whats the current status of mago? Is it still being used to test oneiric? [08:53] As I've seen reports for natty only [10:58] mvo, good morning [10:58] mvo, do you know what this means [10:58] apt-get[27161] trap divide error ip:7f08b489c4ee sp:7fff284854e0 error:0 in libapt-pkg.so.4.11.0[7f08b482a000+114000] [10:59] mvo, I get it with automated ISO testing today. [11:05] jibel: what architecture? [11:05] jibel: any more info? like do you have a log with the commands that ran when it happend? [12:12] mvo, looks like a 64bit specific problem. If you have access to phad, syslog is there: [12:12] http://localhost:8080/job/oneiric-server/ARCH=amd64,HOST=virtual-host,PRESEED=default/46/ [12:12] s/localhost/phad/ :) [12:14] here: http://people.canonical.com/~j-lallement/d-i-syslog.log.gz [12:15] I'll try to reproduce locally because there's not much in the logs. [12:23] thanks jibel, I suspect I know what is going on, I look into this next [12:25] mvo, ok thanks, I trying to get a more useful trace. [12:49] mvo, it is reproducible in a 64bit VM. On i386 there is a dependency failure instead. I'll file a bug. [12:52] jibel: how do you reproduce it? I have one fix that may help, but I'm interessted to see the additional details [12:54] mvo, install oneiric-server-amd64.iso in a VM. The installer crashes when it installs the kernel. [12:57] mvo, bug 823277 [12:58] Launchpad bug 823277 in apt (Ubuntu Oneiric) (and 1 other project) "Ubuntu 64bit failed to install with d-i : apt-get[26978] trap divide error ip:7fc64730d4ee sp:7fffa066b2a0 error:0 in libapt-pkg.so.4.11.0[7fc64729b000+114000] (affects: 1) (heat: 6)" [Critical,Confirmed] https://launchpad.net/bugs/823277 [12:58] it is independent of the VM it fails on kvm and vbox. I haven't tried on bare metal. [13:07] jibel: is there a crash file that you could upload? getting a gdb backtrace of this would rock [13:18] mvo, no crashfile and that's rather difficult to get a trace. When I install the kernel manually in the target environment it doesn't crash. [13:19] hggdh, any idea how we could get a trace for bug 823277 ? [13:19] Launchpad bug 823277 in apt (Ubuntu Oneiric) (and 1 other project) "Ubuntu 64bit failed to install with d-i : apt-get[26978] trap divide error ip:7fc64730d4ee sp:7fffa066b2a0 error:0 in libapt-pkg.so.4.11.0[7fc64729b000+114000] (affects: 1) (heat: 6)" [Critical,Confirmed] https://launchpad.net/bugs/823277 [13:22] jibel: if you run the same comamnd in the shell, does that crash too? [13:23] mvo, nope [13:27] jibel: bad mojo :-( [13:28] mvo, correction, yes it does, I was looking at the wrong place. Ok I think I can get a trace from here. [13:28] good morning hggdh :) [13:29] jibel: good afternoon :-) [13:29] jibel: it only happens on alternate 64? [13:29] hggdh, server too of course [13:30] of course... how could I expect different? [13:30] i386 fails but for a missing dependency on kernel headers, this should autofix with next build. [13:44] jibel: great, let me know once you have somethng that gives more info [13:49] mvo, it seems to crash only when called from debconf-apt-progress . That makes my life a bit more complicated [13:51] jibel: ha! I think I have it [13:52] jibel: i.e. that was the missing clue, its still good to get the trace, but if its a lot of work, we can defer to after I uploaded a fix [13:53] mvo, I feel the pain of debugging in the installer environment, indeed ;) [13:58] heh [13:59] mvo, if there's a build that you think fix this, I can test it before it is uploaded, if it helps. [14:01] jibel: thanks! I uploaded it already, its definitely a required fix, I'm just not sure if its sufficient (but it looks like it should) [14:02] mvo, thanks, I'll keep you informed if it is not sufficient :) [14:03] hehe :) ok === victorp_ is now known as victorp [15:38] mvo, 0.8.16~exp5ubuntu3 fixed the crash but now apt-get hangs when called from debconf-apt-progress. [15:38] mvo, maybe I did something wrong and the testing environment is rather unstable. [15:38] mvo, I'll let you know the results with next images tomorrow. [15:38] mvo, thanks again for the fix [15:40] jibel: great, thanks for verifying, I hope its really the environment, I don't see how the fix could cause it to hang now [15:56] guys, whats the status for mago and tests? [15:57] Is is still active , it worth contributing? [15:57] as mailing list and testsuite branch seems to be inactive since april === yofel_ is now known as yofel