[00:09] Im trying to use EXPECT to write a auto-login script to a telnet session(dont ask me why my work uses telnet). WHat would be wrong with this script [00:10] http://paste.ubuntu.com/881214/ [00:14] philipballew: you may be better served with the perl Net::Telnet module. [00:14] I used expect to do that stuff until I learned about Net::Telnet :) [00:16] greppy, Can you help me with a script for that? [00:17] I dont speak perl [00:17] ah, that could make life a little more interesting, but there are good examples in the docs. [00:17] http://search.cpan.org/~jrogers/Net-Telnet-3.03/lib/Net/Telnet.pm [00:18] philipballew: you can install the module using 'sudo apt-get install libnet-telnet-perl' [00:18] does it matter that I dont know perl greppy [00:18] I've got a bit of a predicament... [00:19] every hour on the 40 my server stops responding for ~30 seconds, and I can't figure out why. [00:19] philipballew: take a look at the docs, there's an example script right at the beginning. [00:19] What, besides cron, should I look at? [00:19] alright. Might be fun to play with [00:20] Rallias: have you taken a look at the logs after the 30 second stall? [00:44] Rallias: it's a cron job [00:44] twb but there's no cron job set to run on the 40 minute mark [00:45] Rallias: you checked /var/spool/cron and /etc/cron* ? [00:46] It may still trigger on the o'clock but take 40 minutes of thinking before its I/O starts [00:48] doesn't have to be cron, but very likely [00:48] oh... [00:49] mysql supports events, so it could be like a cron inside mysql :) [00:49] well wait... I have php fastcgi starting on the 9 and 39, but the download doesn't happen on the 9... [00:49] or anything else that does that, I know wordpress has it's own built in cron too [00:49] by stop responding, you mean web? [00:49] yeah [00:49] then it's probably apache stopping [00:49] but there is still a session in progress [00:49] I run nginx [00:50] well, whatever [00:50] and till that session finishs, it won't stop, so you can start it again [00:50] normally there is a better way to restart that [00:50] but why in the hell would you cron a restart? [00:50] but it knocks out everything as you saw with me disconnecting and reconnecting. [00:52] If shell is still working, you can just ps auxf when the event occurs, and look for D state procs. [00:52] but ssh gets knocked out, so no it doesn't [00:53] Then set up a nohup'd typescript to capture the output every ten seconds, and start running it at half-past the hour [00:53] Then after the event, go in and look at it [00:55] nohup sh -c 'while ps auxf; do sleep 60; done >/tmp/psauxf.log' & [00:55] ...or whatever [00:56] *sigh* [00:56] I hate this thing sometimes :D [00:56] so how long will that script run for? [00:56] Forever [00:56] Obviously once you are past the affected window, you go back and kill it [01:02] so if it is a mysql cron, what should I do to debug it? === baggar11_ is now known as baggar11 [01:03] none of this should cause ssh to die [01:03] unless this is a very very low memory server, and somehow causing everything to oom [01:03] Rallias: turn it off and see if the problem goes away [01:03] Patrickdk: I've seen I/O thrash, too [01:03] I'm running at 250 MB under full most the time [01:04] Rallias: pastebin output of free -m [01:04] i/o thrash won't cause his network connection to drop though [01:04] Patrickdk: well, unless all of apache is in D state [01:04] Patrickdk: but yeah I guess I wouldn't expect it to hang ssh also [01:04] apache yes, but not ssh [01:04] how do I copy with putty? [01:04] never used putty [01:04] twb it even hangs quassel so [01:04] Rallias: NFI [01:06] http://pastebin.ubuntu.com/881267/ [01:09] As far as I can narrow it down, its either nginx, php, or mysql... [01:25] OK so it's probably swap-thrashing rather than OOMing [01:25] Which could conceivably account for stuff like ssh hanging, too [01:25] IME when 2.6 starts to swap you've basically fucked and have to hard-cycle the power to get it to respond [02:20] New bug: #953646 in unixodbc (main) "package odbcinst1debian2 2.2.14p2-5ubuntu3 failed to install/upgrade: ErrorMessage: dependency problems - leaving unconfigured" [Undecided,New] https://launchpad.net/bugs/953646 [02:31] hellop [02:31] for some reason my system boots up without any nameservers in resolv.conf after every reboot [02:31] how do I fix this? [02:32] IP is static [02:33] can I add nameserver under interface config in /etc/network/interfaces ? [02:34] BuenGenio: dns-nameservers 8.8.8.8 8.8.4.4 [02:34] BuenGenio: dns-search mydomaina.com mydomainb.com [02:35] where do I put that? [02:35] BuenGenio: http://www.stgraber.org/2012/02/24/dns-in-ubuntu-12-04/ for more details [02:35] BuenGenio: in /etc/network/interfaces as interface option [02:35] BuenGenio: you can also check "man resolvconf" for some examples [02:36] even after I modify resolv.conf after every reboot it's clean [02:36] yes [02:36] in 12.04 /etc/resolv.conf is managed by resolvconf so if you don't put the right values in /etc/network/interfaces or get your DNS config from a dynamic source, any change to /etc/resolv.conf won't persist after reboot [02:37] and can be lost at any point even if done after boot [02:39] stgraber: resolvconf has been around a long time, and it's optional [02:39] (If precise has forced everyone to install resolvconf, I will be VERY ANNOYED.) [02:42] twb: resolvconf is in ubuntu-minimal in Precise [02:43] Well, FUCK [02:43] ya, changed from optional to required [02:43] twb: per https://blueprints.launchpad.net/ubuntu/+spec/foundations-p-dns-resolving it's possible to turn it off though we recommend people simply use it properly and report bugs so we can fix them (though it looks like we covered most of them already) [02:43] or atleast, could be annoying to remove :) [02:44] Patrickdk: well, removing it is indeed a bit tricky, disabling it is just a matter of making /etc/resolv.conf a regular text file instead of a symlink, though we obviously don't recommend doing that [02:44] stgraber: my problem is I understand how resolv.conf works and I don't want to have some silly extra script layer getting in the way and forcing me to understand it [02:45] Which is incidentally the same objection I had for NM [02:45] twb: then just make it a regular text file and you'll have the old behaviour [02:45] OK, no worries then [02:45] My memory was that it was much more complex than that [02:45] that becomes lots more fun, when you have all kinds of ppp/openvpn/ipsec/... connections going up and down per min on a server :) [02:45] Patrickdk: I do have that [02:46] the problem is, we discovered a lot of scripts and software all of them trying to mess with /etc/resolv.conf most of them overwritting it entirely [02:46] I basically go into (all except one of) their config files and set them to "do not fuck with resolv.conf" [02:46] the only sane way of fixing all that is by forcing them all through a single code path that can deal with merging the duplicate entries and generate a conf file that respects what the libc expects [02:46] stgraber: yeah I understand the purpose of resolvconf [02:46] stgraber: I just didn't want to have to deal with it until I absolutely had to [02:47] the biggest issue that triggered the switch to resolvconf was dhclient's ipv4/ipv6 modes that each fight with the other over who writes /etc/resolv.conf ... that one and read-only boot that was made impossible by having stuff trying to write /etc/resolv.conf :) [02:48] twb: yeah, we made sure that the conversion of /etc/resolv.conf to a symlink only happens the first time resolvconf gets installed, not at boot time and not on upgrades. So converting it back to a regular file basically tells resolvconf to stop doing anything (and it should also prevent anything that's resolvconf-aware from touching your /etc/resolv.conf) [02:49] Ah, dual-stack dhclient would be an issue, yeah [03:02] I've got a quick question about cron [03:02] how would I get a job to run once every 25 hours? [03:03] anyone using spamassassin here? [03:03] Rallias: run it every hour, check if the hours since epoch is divisible by 25, if not, exit 0 [03:04] programiming has never been my strong suit... [03:08] That's not programming, that's basic sysadmin work [03:08] Something like $(($(date +%s) / 3600 % 25) === fenris is now known as Guest47264 [04:26] hi [04:26] how to see loaded modules in startup? [04:27] i found one error but I dont know how to scroll up [04:28] i found one error but I dont know how to scroll up [04:28] i found one error but I dont know how to scroll up [04:42] linocisco, to scroll through the screen buffer on the console, use shift+pgup and shift+pgdn [04:48] qman__, not working [05:14] qman__: that will scroll through video buffer, and if he has chvt'd it will be empty [05:14] scrolling in screen is achieved with ^A^[ to enter copy mode, then vi movement commands (^B,^F,hjkl). Leave it with ^G. [06:04] seekwill: What have you broken now? [06:04] I didn't break anything! [06:05] Not yet, at least [06:14] what happens when support for the ubuntu distro im using expires...just no more updates ? [06:14] wooter: you get nagged about it every time you log in [06:14] ah thanks [06:15] Note that different packages expire at different times [06:15] will things like apache still update ? [06:15] So even though you theoretically have five years of support, most of your packages only get 18mo or 3y [06:15] wooter: no, once the release is EOLd, none of it gets security updates [06:16] Ideally you should plan migration to the next LTS release several months before your current release is EOLd [06:16] ah k..my server is 10.10, is up next month [06:17] LTS comes out as same time as this expires [06:18] THat is because 10.10 is not an LTS release [06:18] You should probably sit on 10.10 until 12.04.1 is release, then migrate to that. [06:18] yeah, i wasn't thinking when i installed it [06:18] Since it's only a month or two, it's probably not worth the grief of trying to migrate to 11.04 until then [06:19] Unless this is a front-line server, in which case security might be a bigger issue [06:19] nah its not front line [06:19] still, it took alot of man hours to get it all configured how it is now [06:19] dopnt want to do it again for 5 years [06:20] wooter: upgrading doesn't mean reinstalling from scratch [06:20] This is not a cowboy outfit like RHEL [06:22] yeah, ive got a few hardware and platform specific scripts going on, everytime it updates, stuff happens like entire sound sub system changes or something like that [06:22] ill try upgrade first for a laugh [06:22] !upgrade [06:22] For upgrading, see the instructions at https://help.ubuntu.com/community/UpgradeNotes - see also http://www.ubuntu.com/desktop/get-ubuntu/upgrade [06:23] Please read the documentation first [06:25] thanks twb [06:31] qman__, shift pgup and shiftpgdwn not working [06:33] Hello, on current precise grub does not set a timer when the previous boot failed, is it intended? [06:33] linocisco: 16:14 scrolling in screen is achieved with ^A^[ to enter copy mode, then vi movement commands (^B,^F,hjkl). Leave it with ^G. [06:34] twb, i m now talking about on CLI . not on VI [06:34] linocisco: I'm talking about GNU Screen. [06:35] twb, it also doesn't work [06:35] twb, my server is on VM [06:35] twb, as guest [06:36] linocisco: dunno then, sorry [06:36] twb, thanks anyway [06:47] hi all [06:48] regarding squid, i have WAN card with DHCP with authentication to get internet and, LAN card which has static IP. when I run "nmap WAN DHCP's IP", I got 3128 squid-http only [06:48] other ports are not open [06:49] is that the reason why my client does not get internet through squid? [06:50] hi all, i m in the middle of tasksel, now installing package 1%, I want to stop it with Ctrl+C, so how could I do to properly stop? [06:51] hi all [06:52] how to interrupt of installation package using tasksel? === rye_ is now known as rye === aarcane is now known as root________ [08:54] what do I need to do to increse the "max open files" limit for mysql (I don't mean the user but mysqld as I have no idea where its imposed limits come from) [09:12] * bluefrog test [09:12] pouet [09:26] a_ok: take a look at the upstart configuration in /etc/init/mysql.conf [09:26] a_ok: and man 5 init [09:27] you can set the max files limit with an override [09:27] jodh: hello, I am stalking you regarding https://launchpadlibrarian.net/96606883/lp935585.jpg (bug #935585) which is currently being displayed on my screen and I don't want to boot it until I find out how to debug the reason [09:27] Launchpad bug 935585 in upstart "[kernel panic] init: log.c:786: Assertion failed in log_clear_unflushed: log->remote_closed" [High,Confirmed] https://launchpad.net/bugs/935585 [09:32] rye: the work around is documented on the bug here: https://bugs.launchpad.net/ubuntu/+source/upstart/+bug/935585/comments/4. I've made the code changes for a proper fix and will soon be advertising an updated Upstart package (in a PPA) for initial testing. As such, we would welcome feedback from yourself and as many others as possible. I'll update the bug when the PPA version is available. [09:32] Launchpad bug 935585 in upstart "[kernel panic] init: log.c:786: Assertion failed in log_clear_unflushed: log->remote_closed" [High,Confirmed] [09:36] jodh: umm, yes, i am aware of the workaround. Since this is not a production machine i thought I can assist in debugging it, since as I understand the reason for such behavior was not found === Kiall_ is now known as Kiall [10:11] New bug: #953876 in exim4 (main) "Can not specify different DNSBLs for IPv4/IPv6" [Undecided,New] https://launchpad.net/bugs/953876 [10:25] c7usi [10:25] hey guys, is there a simple way to determine which flags the apt-get version of nginx is compiled with? [10:27] <_ruben> unless nginx can provide that info itself, you'll likely have to inspect the source package [10:28] cool, looks like I was on the right track. am downloading it now and will inspect the `debian/rules` file [10:38] Hi [10:39] jamespage: just saw your comment on #928501, going to prepare it for the FFe now [10:39] Can I take MySQL Debian package to install 5.5.21 version on Ubuntu 11.10 ? Or is it recommended to build from source? [10:43] jodh: ok, so in order to move along - there is no need to debug the kernel panic issue on startup, right? I.e. I can proceed with booting it normally? [10:46] New bug: #953899 in unixodbc (main) "package odbcinst1debian2 2.2.14p2-5ubuntu3 failed to install/upgrade: ErrorMessage: dependency problems - leaving unconfigured" [Undecided,New] https://launchpad.net/bugs/953899 [10:46] New bug: #953901 in unixodbc (main) "package odbcinst1debian2 2.2.14p2-5ubuntu3 failed to install/upgrade: ErrorMessage: dependency problems - leaving unconfigured" [Undecided,New] https://launchpad.net/bugs/953901 [10:49] rye: I'm not sure I understand what you are saying. What would be very useful is if you would try the PPA in #11 on that bug and comment on the bug. [10:52] rye: We were not able to reproduce the problem as reported, but I was able to identify the fix for what is essentially a timing issue (how quickly the log partition becomes writeable). [10:54] jodh: i meant that I have the server with the error right here, so i might have altered the packages to add some debugging output that might have helped. But the "how quickly the log partition becomes writeable" seems correct, however I am still not sure why this can be 100% triggered when fsck is forced to run. Ok, will add the ppa then [11:04] it looks like the mysql upgrade from last night is breaking permissions.. anyone else having troubles? [11:07] ok, ppa version for upstart works properly [11:12] bencer: I'll see if I can track down an archive admin re native packaging but don't let that hold you up on the FFe [11:13] should i remove that from the ffe then? [11:18] New bug: #953915 in cloud-init (main) "Missing .deb dependency on python-oauth" [Undecided,New] https://launchpad.net/bugs/953915 [11:33] hrmf! [11:33] anyone here using automount? [11:33] I've setup auto.master pointing to auto.home and when logging in, nothing happens, automount doesn't output a single line of debug info [11:36] cd /home/rsk works - logging in doesn't trigger mounting [11:43] (anyone?) [11:47] RoyK: that sounds most odd [11:48] RoyK: when you login and it doesn't mount, is your expected home dir there ? [12:06] do openstack issues get discussed on a particular ubuntu mailing list, or is it left to the main openstack mailing list? [12:15] ikonia: sec [12:19] hm... it seems if the user is in nis, the homedir isn't fetched on login and automount doesn't do much [12:19] if user exists in local files, it works (somewhat) [12:36] hmm [12:37] been awhile since I used automount, it was always unreliable for me [12:37] pam_mount worked well for login mounts though [12:41] New bug: #953985 in keystone (universe) "keystone does not attempt to reconnect to sql server" [Undecided,New] https://launchpad.net/bugs/953985 [13:38] RoyK: sorry for the delay, but do you have nis setup to point at the automounter, eg: auto.master (auto_master) and auto.home (auto_home) [13:40] ikonia: it worked out - lots of mess in this setup [13:41] RoyK: care to expand ? [13:41] was it the nis auto.home map (for example) === Ursinha` is now known as Ursinha [13:58] bencer: just got an ack on the FFe and an archive admin to review once they land in NEW [14:16] New bug: #954068 in euca2ools (main) "describe-images omits empty fields platform and root_device_type" [Undecided,New] https://launchpad.net/bugs/954068 [14:18] ikonia: no I don't care to explain ;) [14:26] fair enough [14:38] Preference? nova-database-mysql or nova-backend-mysql; this is for dbconfig-common support [14:42] I'm looking for something rsync-like which runs as a daemon and pushes changes automatically, any advice? [14:43] As I understand it, drbd waits until the write has been confirmed on both hosts before considering the file "written", which is not what I want. [14:50] chmac: rsync can run as a daemon [14:50] guess, you want the automatic part [14:51] pabelanger: Does it monitor files and push those changes automatically? Might well do what I'm after... [14:53] chmac: no, I don't think so [14:53] you'd have to read up on it [14:54] pabelanger: Yeah, I think it acts as a daemon to receive rsync connections from other hosts, rather than monitoring the filesystem. [14:55] pabelanger: I found unison and lsyncd which might do what I'm after [14:55] pabelanger: I essentially want something like rsync running ever 5 minutes from cron, but slightly faster and with less overhead :-) [14:59] pabelanger: Thanks for your input btw [14:59] np [15:10] bencer: around? need to talk about ebox->zentyal transition [15:21] New bug: #954127 in dhcp3 (main) "Network configuration fails during installation of 12.04" [Undecided,New] https://launchpad.net/bugs/954127 [15:22] I need a kernel module which only seems to exist in 2.6.35-28-generic - where can I download just the module to compile it against my current 2.6.32-28 kernel? [15:24] SpamapS: hi! I see you did some stuff with the txzookeeper test suite. what do I need to do to run it? I did: cp -a /usr/share/pyshared/txzookeeper /tmp ; cd /tmp/txzookeeper ; trial txzookeeper [15:24] iclebyte: look inside the 2.6.35-28-generic kernel sourcecode [15:25] the ubuntu specific source or direct from kernel.org? [15:25] SpamapS: that mostly works except for: test_session [ERROR] [15:25] exceptions.AssertionError: ZOOKEEPER_PATH environment variable must be defined [15:25] iclebyte: the ubuntu specific source [15:26] jdstrand: you need a running zookeeper [15:26] SpamapS: I installed zookeeperd [15:26] New bug: #912508 in txzookeeper "test suite can wipe out existing zookeeper on local box: should Configure zk server as part of test fixture" [Medium,Confirmed] https://launchpad.net/bugs/912508 [15:26] jdstrand: wait no, I think that one starts zookeeper up.. ZOOKEEPER_PATH=/usr/share/java/zookeeper.jar might work [15:27] jdstrand: hard to recall.. one of the test suites trashes a running zookeeper [15:28] SpamapS: I saw that bug and actually just commented in it. I am writing a qrt script so I can remember how to do the test suite in case it doesn't get enabled in the build. these run in vms so I don't care about trashing things [15:28] iclebyte: you can also google for the module name and try to locate any upstream source repository [15:28] and then build the module standalone [15:28] yea it's a Qlogic one. having a hard time finding it [15:29] it's called qlcnic [15:29] 10gigE fibre card [15:29] SpamapS: this is it: ZOOKEEPER_PATH=/usr/share/java/ trial txzookeeper [15:30] SpamapS: perfect, thanks! [15:30] PASSED (successes=136) [15:30] jdstrand: the reason I didn't want to run the test suite in the build if it trashes zookeeper is that some people build in chroots on their machines.. would be a shame to accidentally hose somebody's zookeeper. [15:31] jdstrand: I think it will be relatively easy to modify the suite to start a bastion zookeeper and run the tests against that === Lcawte|Away is now known as Lcawte [15:31] SpamapS: no, I get that. It would be nice to start zookeeperd on a different port or something [15:31] * jdstrand nods [15:32] jdstrand: we do that for juju.. so we can even steal that code, maybe even push it back into python-zookeeper as a utility class. [15:33] cool, yeah [15:33] iclebyte: the module source got merged into the upstream linux kernel and sent in by qlogic themself. its maintained in the git tree at kernel.org [15:34] jdstrand: thanks for the zookeeper review. Next step is: update man pages to mention network security options. Anything else? [15:34] iclebyte: this was the initial commit http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=af19b49152bdb68fda894183e88096d6d1aa5c3d [15:34] SpamapS: nope, I think that should be it [15:34] i have since then moved around inside the kernel tree [15:37] xranby, can this git interface generate a tarball with all the contents or do I need to download each file manually? [15:39] SpamapS: oh, you might check out the deprecation warning [15:39] iclebyte: take this patch http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=patch;h=af19b49152bdb68fda894183e88096d6d1aa5c3d [15:39] iclebyte: and apply it on your kernel tree [15:39] SpamapS: it isn't required for main inclusion; just fyi [15:39] jdstrand: I will.. I'd be surprised if upstream hasn't addressed it.. they have been pretty on top of such things which is why there aren't many warnings. [15:39] does it not need to be for a specific kernel source version? [15:40] iclebyte: that all depends on if the code itself uses features only found in newer kernels [15:40] you have to read and test the code yourself on your old kernel [15:40] the code itself do not care which kernel version its used on [15:41] I thought diff would because it references specific lines in the code [15:41] this are the initial commit so its all new files [15:41] 8 files changed, 8728 insertions(+), 0 deletions(-) [15:43] jdstrand: fyi: https://issues.apache.org/jira/browse/ZOOKEEPER-909 .. will open a bug report which we can close when 3.4.0 is uploaded [15:43] jdstrand: (thats the deprecation warnings) :) [15:44] SpamapS: oh, we are getting 3.4.0 in precise? [15:44] * jdstrand was hoping we would for the kerberos bits [15:44] jdstrand: no [15:44] jdstrand: btw, that brings up a different issue. PHP 5.4 .. it looks like Suhosin has not shipped for 5.4.0 yet [15:44] mdeslaur: ^^ [15:45] micahg: ^^ [15:45] jdstrand: perhaps we should discuss this at your team meeting, but I feel that its more your team's call than mine. For server users, 5.4.0 is by far the preferred choice, but w/o Suhosin.. it puts a larger burden on you guys [15:45] SpamapS: well then, looks like we aren't getting 5.4 after all :) [15:45] jdstrand, 3.4.x is still considered a beta release by upstream [15:46] I see [15:46] I asked suhosin's authors for an update.. but have not heard anything [15:46] New bug: #953258 in juju "juju executing hooks with umask of 0077" [Undecided,Fix released] https://launchpad.net/bugs/953258 [15:46] how do I apply this patch to a kernel source tree? [15:46] SpamapS: ah, yeah, I was about to ask that [15:47] headers are in usr/src/linux-headers-2.6.32-28 [15:50] ? [15:55] hello guys, i'm on 11.04 server, i've tried to change hostname by editing /etc/hostname & /etc/hosts but after reboot it's still the same..any idea? [15:56] alex88, did it save okay? [15:56] New bug: #954160 in zookeeper (universe) "deprecation warnings during build: org.jboss.netty.channel.ChannelPipelineCoverage" [Low,Triaged] https://launchpad.net/bugs/954160 [15:56] iclebyte: yup.. wait.. if it's important, it's a openvz container [15:56] not played with openvz so I can't comment. [15:56] let me check also that… [15:57] iclebyte: yup, that was the problem..thank you anyway man ;) [15:58] welcome. [15:58] smoser, no hvm ami's anywhere but east-1? [15:58] jesus, running Maverick is going to be easier than this. [15:58] no hvm anywhere but us-east-1, jcastro [16:06] iclebyte, what are you doing? [16:14] cwillu_at_work, I need a driver which only exists in 2.6.35 on my ubuntu 10.04lts server system [16:14] I have the patch http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=patch;h=af19b49152bdb68fda894183e88096d6d1aa5c3d [16:14] iclebyte, you aware that newer kernels will work fine on it? [16:15] i.e, you should be able to install a 11.10 kernel deb without issue [16:15] I was thinking that maybe there is another repo I can add then just upgrade my kernel. exactly where to find such a repo i'm not sure === cmagina_ is now known as cmagina [16:25] smoser: hey-- I saw that you packaged txaws. what is aws-status supposed to actually do? [16:26] its a gui i think , right ? [16:26] i ran it at once [16:26] s/at once/once/ [16:26] it was long ago [16:27] there must be a backported kernel package somewhere? === jussi01 is now known as jussi [16:30] smoser: well, it prompted me for some credentials, then didn't do anything [16:30] * jdstrand moves along === larva_ is now known as larva [16:59] jamespage: i'm around, let me know what you need about the transition [17:00] bencer, hey - so I updated the bug report as well [17:00] let me check first then [17:01] bencer: as the packages stand now if someone upgrades a system with ebox it won't automatically get replaced with the equiv zentyal packages [17:01] this was pointed out to me by an AA [17:01] jamespage: let me put you in context [17:02] current ebox packages are totally broken in oneiric, and in a pretty bad shape in lucid [17:02] so is unlikely anybody is using them :) [17:02] also the configuration backend changed [17:02] lucid->precise is quite a likely upgrade path tho [17:02] so you need a migration tool to change the configuration format [17:03] anyway, the configuration paths are different [17:03] so no configuration is overwritten [17:05] SpamapS, around - would appreciate your input into this conversation ^^ [17:06] jamespage: they story is that ebox packages in the past were not really maintained, were uploaded and forget about them [17:07] now is one of the goals of the zentyal project to have working packages in ubuntu [17:07] so we will put efforts on their maintenance [17:07] jamespage: looking [17:08] bencer: we should still make sure that zentyal causes the removal of ebox so there are no problems. [17:09] SpamapS, bencer: well the packages as is cause a removal - but its not automatic - you have to switch to zentyal [17:10] yes, if you do apt-get install zentyal, ebox packages are removed [17:10] how we should make the upgrade path then? [17:11] with a dummy package? [17:11] Ah, there needs to be a transitional, empty ebox package [17:12] bencer, see http://wiki.debian.org/Renaming_a_Package [17:12] yes exactly, just an ebox package, which depends on zentyal [17:12] it can be removed in 12.10 [17:13] bencer: we agree that having old broken ebox around is unlikely.. but if it was around, and causing problems for the user.. they'll be quite happy when its replaced w/ Zentyal [17:13] yup, i agree [17:13] and shoud be easy to do [17:14] jamespage: SpamapS where should be do the dummy package [17:14] bencer: it forms part of the source package that will replace it [17:14] on ebox or on libebox, so we can set it to oldlibs and be automatically suggested to be removed? [17:15] bencer: zentyal-common replaces libebox so the dummy package should be provided by it [17:16] jamespage: currently we have the replaces header in all packages [17:17] bencer: yes - but it needs a transitional package as well [17:17] they way to go is to still maintain the replaces and add the dummy package for libebox, right? [17:17] yep - each package should have a Breaks/Replaces and a transitional package for the package it replaces. [17:18] cool, were do you want us to push these changes? [17:18] jamespage: a provides isn't strong enough.. [17:18] you need a dummy package [17:18] on our ppa? or in the bzr import you created? [17:18] oh [17:18] sorry I misunderstood [17:19] PPA is fine [17:19] bencer: I really don't want to make any changes to the packages you provide if possible [17:20] fair enough [17:20] going to talk with jacalvo and we will push that changes and i'll update back you when ready [17:22] bencer: do you want me to check before you upload to PPA? [17:22] that way we avoid another cycle of review if there are any issues [17:22] jamespage: i'll send you a link to our git changeset when ready [17:22] bencer: great! [17:28] SpamapS, utlemming says that you said you would do: [17:28] https://bugs.launchpad.net/ubuntu/+source/apt/+bug/948461 [17:28] Launchpad bug 948461 in apt "apt-get hashsum/size mismatch because s3 mirrors don't support http pipelining correctly" [High,Confirmed] [17:28] as SRUL [17:29] SpamapS, reboot testing? [17:30] jamespage: yes! [17:30] jamespage: whats the word? [17:30] SpamapS, so - I did some digging as gema asked me to take a look at the failures [17:30] smoser: I'll take a look at it yes [17:30] and I knew you where at charm school [17:30] I discovered three different bugs! [17:31] smoser: best if somebody else uploads, so I can just do it as an SRU review [17:31] jamespage: only 3? :) [17:31] SpamapS, bug 953081 [17:31] Launchpad bug 953081 in php5 "libapache2-mod-php5 postinst script does not enable module" [Undecided,New] https://launchpad.net/bugs/953081 [17:31] ah. yeah. thats fine. i thought utlemming suggested that you were going to backport fix and upload ... [17:31] utlemming, ^ [17:31] SpamapS, bug 953144 [17:32] Launchpad bug 953144 in ubuntu-server-iso-testing "Post-reboot tests are not correctly injected into test-case" [Undecided,New] https://launchpad.net/bugs/953144 [17:32] and one in Jenkins [17:32] SpamapS, basically the standard lamp test suite is being run twice - it fails pre-reboot because of that first bug [17:32] haha nice [17:32] smoser: yeah, I'll do the backporting work [17:33] SpamapS: it then runs it again due to the second bug post reboot [17:33] but it does pass [17:33] but Jenkins then borks as it gets confused because the test suite has the same name [17:33] and reports a double failure from the first run [17:34] I have a local fix for usit that I'm working on ATM - had to put it down today for other stuff [17:34] jamespage: ok, so that makes sense because I never actually ran the test as part of jenkins [17:34]