[09:52] cjwatson: Raring server installations fail at the clock setup step with today's images, reported bug 1153992 for that [09:52] Launchpad bug 1153992 in debian-installer (Ubuntu) "Raring server installations fail at the clock configuring step" [Undecided,New] https://launchpad.net/bugs/1153992 [09:55] cjwatson: infinity: just noticed that this affects precise d-i installations as well, I'll update the ^bug accordingly [10:07] psivaa: Did that machine fail to DHCP? [10:07] psivaa: Or did the DHCP server change in some curious way? [10:09] infinity: if you mean the host machine, then i dont think so because the failure is occurred in a vm on my local host machine and in our server [10:10] Hrm, no, maybe not related to DHCP... [10:10] But I can't see how this would have changed in both precise and raring recently. Did your testing methodology change? [10:11] The two packages involved here haven't changed in precise since well before release. [10:11] infinity: not really, i was able to reproduce this manually too, but what do you think about this entry in the log [10:11] Mar 12 07:18:20 debconf: <-- 10 netcfg/dhcp_ntp_servers doesn't exist [10:11] That was the red herring I was looking at. But it just means that the DHCP server didn't give any NTP info, which is perfectly valid. [10:11] And is what I would see here too. [10:12] And what most networks would see. [10:12] But it then looks like it's trying to display a question at critical priority that you skip? [10:12] Are you running fully preseeded/noninteractive and obtusely skipping that question? :P [10:14] d-i clock-setup/ntp boolean true [10:14] d-i clock-setup/ntp-server string ntp.ubuntu.com [10:14] ? [10:15] Though, the second one should have a sane default. [10:15] And does. [10:15] Actually, both do. [10:16] Mar 12 07:18:20 debconf: --> INPUT medium clock-setup/ntp [10:16] Mar 12 07:18:20 debconf: <-- 30 question skipped [10:16] Mar 12 07:18:20 debconf: --> GO [10:16] ^-- Are you skipping it somehow in your preseed? [10:16] infinity: nope, the clock is normally set from the location and also it does not explain why it should happen in my manual installation [10:17] also the same preseeds worked fine till yesterday [10:18] infinity: d-i clock-setup/utc boolean true [10:18] d-i clock-setup/utc-auto boolean true [10:18] d-i time/zone string US/Eastern [10:19] are the entries in the preseed [10:21] Curious. I may leave this one to cjwatson, since it's 4:20am here. [10:22] infinity: ok :), thanks for looking into it though [10:52] daylight saving time change? [10:52] tz-data change in raring? [10:53] nothing else looks obvious. [10:56] Hmm [10:56] netcfg/dhcp_ntp_servers not existing is *not* a red herring - the db_get there has no || true, so a template existence failure will crash clock-setup [10:57] I suspect a regression from my cdimage changes yesterday [10:59] psivaa: Thanks - I'll take it from here [10:59] cjwatson: thank you [10:59] Indeed, netcfg is on the 20130311 raring server images but not 20130312 [10:59] WHOOPS [11:05] convergence ! ... just telll them to use 3G :P [11:24] cjwatson: Oh, hah. Did that also break precise? [11:25] I expect so [11:25] cjwatson: I misread it as "the template doesn't exist because dhcp didn't give us results", but I suppose the templates there regardless, just without a value. [11:25] But I'm not caffeinated enough to have figured it out yet [11:25] indeed, 10 doesn't exist vs. 0 OK RET="" [11:25] This is why I shouldn't debug at 4am, though, and why I handed it off. :P === kentb-out is now known as kentb === ogra_` is now known as ogra_ === kentb is now known as kentb-afk === kentb-afk is now known as kentb === ev_ is now known as ev === maxb_ is now known as maxb