| RoAkSoAx | \sh, awesome, I was planning to look into that to try to get it for Maverick | 00:00 |
|---|---|---|
| RoAkSoAx | mathiaz, what kind of other alternatives? | 00:00 |
| mathiaz | RoAkSoAx: well - whatever you need to do to register a puppet client | 00:01 |
| mathiaz | RoAkSoAx: I don't have a specific example | 00:01 |
| mathiaz | RoAkSoAx: as I would use the cloud-config support for puppet | 00:01 |
| RoAkSoAx | mathiaz, I see then... i'll look into it then. And btw.. would that work with regular KVM's? or Only for the Cloud instances? | 00:03 |
| mathiaz | RoAkSoAx: only UEC/EC2 images | 00:03 |
| mathiaz | RoAkSoAx: it also depends what you use to run your kvm | 00:03 |
| mathiaz | RoAkSoAx: for now the necessary information is brought via the user-data option | 00:04 |
| RoAkSoAx | mathiaz, well I would like to test it on my local machine without having to implement a UEC | 00:05 |
| mathiaz | RoAkSoAx: it == ? | 00:05 |
| RoAkSoAx | mathiaz, ?? :/ | 00:06 |
| RoAkSoAx | what do you mean | 00:06 |
| mathiaz | RoAkSoAx: you wanna test *what* without the cloud? | 00:07 |
| RoAkSoAx | mathiaz, i mean auto registration of the puppet client | 00:08 |
| RoAkSoAx | as if they were running in the cloud | 00:08 |
| RoAkSoAx | but they would be running on a local machine on KVM | 00:08 |
| mathiaz | RoAkSoAx: well - that wouldn't work since the work done is to enable automatic registration in the cloud | 00:09 |
| RoAkSoAx | mathiaz, that is managed by puppet... correct? | 00:10 |
| mathiaz | RoAkSoAx: the cloud? no it's not managed by puppet | 00:11 |
| RoAkSoAx | mathiaz, no i the integration of puppet to EC2/UEC will allow a puppet server to register to a puppet master that are both running in the cloud? | 00:12 |
| RoAkSoAx | s/puppet server/puppet client | 00:13 |
| RoAkSoAx | which also means that the instance will be registering to the cloud, correct? | 00:13 |
| mathiaz | RoAkSoAx: well - it's the other way around | 00:16 |
| mathiaz | RoAkSoAx: the cloud starts an instance | 00:17 |
| mathiaz | RoAkSoAx: and the instance will be able to register to any puppet master | 00:17 |
| mathiaz | RoAkSoAx: I need to jet out | 00:17 |
| mathiaz | RoAkSoAx: I'd suggest to read up the series of blog post I made about puppet-UEC/EC2 integration | 00:17 |
| RoAkSoAx | mathiaz, i will thanks :) | 00:18 |
| mathiaz | RoAkSoAx: http://ubuntumathiaz.wordpress.com | 00:18 |
| smoser | debug6 hggdh | 00:19 |
| smoser | hggdh, i've never seen that error. sigh. | 00:20 |
| hggdh | smoser: heh. which error? ;-) | 00:23 |
| smoser | red only filesystem | 00:23 |
| smoser | read even | 00:23 |
| hggdh | smoser: yeah, this sucks... I just opened bug 565101 about that | 00:24 |
| uvirtbot | Launchpad bug 565101 in eucalyptus "walrus reports java.lang.OutOfMemoryError: Java heap space" [High,In progress] https://launchpad.net/bugs/565101 | 00:24 |
| hggdh | smoser: could you prepare a debug6-compatible image for the RC? | 00:24 |
| smoser | hggdh, you mean take the rc image and enable some debug ? | 00:25 |
| smoser | heres the problem with that. | 00:25 |
| smoser | bug 566793 | 00:25 |
| uvirtbot | Launchpad bug 566793 in eucalyptus "euca-get-console-output gives first 64k of output, not most recent" [Low,New] https://launchpad.net/bugs/566793 | 00:25 |
| smoser | the reason that is a problem is that if you turn upstart debugging on, output is so verbose that you don't get the end of it | 00:26 |
| smoser | i worked around that here by ssh'ing to the NC and getting the console log from there, but in your situation with multiple NC, it'd be harder | 00:27 |
| \sh | hggdh, is tomcat somehow involved? | 00:27 |
| hggdh | \sh I do not think so | 00:31 |
| hggdh | smoser: yes, have the debug code on the RC -- but only if we will learn something new from it, I guess | 00:31 |
| hggdh | damm, gave the wrong bug # on the console thingy | 00:32 |
| hggdh | the readonly FS is bug 567592 | 00:32 |
| uvirtbot | Launchpad bug 567592 in cloud-init "rm: cannot remove `/var/lib/urandom/random-seed': Read-only file system" [Undecided,New] https://launchpad.net/bugs/567592 | 00:32 |
| hggdh | \sh no tomcat (just checked) | 00:33 |
| \sh | hggdh, well, the bug sounded like a long time hanging tomcat bug regarding "re-starting/re-deploying" tomcat webapps.. | 00:34 |
| smoser | hggdh, ok. i have registered 3 amis now | 00:39 |
| smoser | euca-describe-images | grep "emi.*20100420" | 00:39 |
| smoser | lucid-20100420-debug-upstart/lucid-server-uec-amd64.img.manifest.xml | 00:39 |
| smoser | lucid-20100420/lucid-server-uec-amd64.img.manifest.xml | 00:39 |
| smoser | lucid-20100420-debug-upstart-loud/lucid-server-uec-amd64.img.manifest.xml | 00:40 |
| smoser | debug-upstart-loud has debug at 'info' | 00:40 |
| smoser | debug-upstart actually has no upstart debug, but does have some from cloud-init | 00:40 |
| hggdh | smoser: thank you. I will see what happens with them ;-) | 00:49 |
| smoser | but as i said, hggdh, due to that console output bug, if you don't capture the whole console output, its not that useful | 00:50 |
| smoser | but for this read-only one, i have mountall debug on | 00:51 |
| smoser | so that might help | 00:51 |
| smoser | ah. wait. | 00:51 |
| smoser | we've seen this before | 00:51 |
| smoser | the plymouth not available | 00:51 |
| smoser | kirkland was moaning about this in -devel the other day | 00:51 |
| smoser | kirkland, did you open an issie on this ? | 00:52 |
| hggdh | smoser: as far as I can remember pretty much all failed instances (I took out the console output for success) seem to have a plymouth barfing | 00:58 |
| smoser | yes, but in this case mountall decided to do something about it :) | 00:58 |
| smoser | thats the difference | 00:58 |
| hggdh | heh | 00:58 |
| hggdh | which image has mountall at debug? | 00:58 |
| smoser | both | 00:59 |
| smoser | both the debug ones | 00:59 |
| smoser | the -loud has upstart to | 00:59 |
| smoser | stick with the non loud | 00:59 |
| hggdh | yes, sounds like a very good idea. | 01:00 |
| hggdh | there we go :-) | 01:00 |
| smoser | i just instaleld the -debug0 package that i put together for your other bug | 01:00 |
| smoser | hggdh, i have to run | 01:01 |
| smoser | sorry | 01:01 |
| === sferry is now known as sammah | ||
| hggdh | smoser: thank you, and have a nice evening | 01:03 |
| === sammah is now known as samferry | ||
| ajsie | anyone here? | 02:24 |
| domas | hi! | 03:11 |
| domas | anyone seeing apparmor related oops on install ? http://p.defau.lt/?nBDkcPAlmAk_DmALqzZzag | 03:11 |
| sbeattie | domas: I haven't seen anything like that, but jjohansen will be interested. | 03:13 |
| domas | I'm having troubles running installer on 2x6 core opteron boxes | 03:13 |
| domas | thats brand new CPU, so may be related :) | 03:13 |
| domas | though managed to install one box :) | 03:14 |
| jjohansen | domas: can you file that in a bug | 03:14 |
| domas | have to figure out how can I contact canonical support :)) | 03:15 |
| jjohansen | domas: also which kernel? | 03:15 |
| domas | 2.6.31-14-generic #48-Ubuntu | 03:15 |
| jjohansen | what are you trying to install? | 03:15 |
| domas | different oops now: http://p.defau.lt/?FSq_675qXqvJxQDZLdwATw | 03:15 |
| domas | karmic | 03:16 |
| domas | this is unattended install, that succeeds on all our other machines | 03:16 |
| domas | just this class of machines fails | 03:16 |
| jjohansen | domas: hrmm, strange | 03:17 |
| domas | I installed one :) | 03:17 |
| domas | tried karmic, failed, tried hardy ,failed, succeeded with next karmic install :) | 03:18 |
| jjohansen | domas: have you tried Lucid? | 03:18 |
| domas | I'm not sure I have network image for that one around | 03:18 |
| domas | is it released yet? | 03:19 |
| domas | ah, few more weeks | 03:19 |
| jjohansen | domas: its RC so very close, its kernel would probably better support that rig | 03:20 |
| MagicFab | domas, there are dailies. | 03:21 |
| domas | anyway | 03:23 |
| domas | pure magic is that I have it installed on one box ;-) | 03:23 |
| domas | let me retrace the steps ;-))) | 03:23 |
| domas | hardy kernel was locking up on all older machines of this type too :) | 03:28 |
| domas | hmmm, now hardy installer is working | 03:30 |
| domas | btw, is there a way to create stripe-aligned (or 1MB-aligned) partitions with automated installer? :) | 03:32 |
| jjohansen | domas: not that I know of, but then I don't play with that stuff much | 03:43 |
| bluethundr | I've specified a directory on my new mail server that I would like to mount separately from the rest of the file system (along with a couple of others that do work correctly). I am specifying the directory as this in fstab: /dev/sdm /var/spool/mail/virtual ext3 defaults,noatime,noacl,data=ordered 1 2…. it shows up as /var/mail/virtual …. http://pastebin.com/LmvZBMWr | 04:05 |
| deslector | hi, what is a good way to monitor a server load for a period of time? | 04:38 |
| * deslector wonders if it is possible to monitor power consumption too | 04:42 | |
| domas | you can monitor power consumption, if your chassis supports that :) | 04:42 |
| maxagaz | hi | 04:42 |
| domas | hehe | 04:42 |
| domas | actual_power = 300 | 04:43 |
| maxagaz | when sending a message with the command mail, can I set the sender's address ? | 04:43 |
| domas | maxagaz: use 'sendmail' and provide headers | 04:43 |
| deslector | domas, hmm... I don't know, but don't think so... this is one of the cheap ProLiants | 04:48 |
| deslector | I'll check into it, though... | 04:48 |
| deslector | hmm... pwrkap may be what I want | 04:51 |
| uvirtbot | New bug: #567701 in ntp (main) "Date not updated by ntp when network interfaces started" [Undecided,New] https://launchpad.net/bugs/567701 | 05:36 |
| swift | guys, I have two NIC interface configured on different subnets on my ubuntu server | 08:03 |
| swift | one on 192.168.44.x and the other on 192.168.45.x | 08:03 |
| swift | please confirm if the gateway used by these interfaces should be different | 08:04 |
| swift | when i do a 'route -n' , i see that the gateway for both is 0.0.0.0... which then is 192.168.44.2 | 08:04 |
| swift | so.. is this a problem? | 08:05 |
| swift | please advise | 08:05 |
| persia | That entirely depends on the network configuration. | 08:05 |
| persia | But unless you're doing something especially fancy, there should only be one route to 0.0.0.0 on a host. | 08:05 |
| swift | persia... how can a 192.168.45.x subnet us a 192.168.44.2 gateway? | 08:05 |
| swift | persia... shouldn't 192.168.45.x use itself is the gateway? | 08:06 |
| persia | Because your host routes packets from 192.168.45.x to the 192.168.44.x interface to go to the 192.168.44.2 device, which then routes them somewhere else. | 08:06 |
| persia | No. | 08:06 |
| swift | persia.. that's what I dont want.. I done want packets to go from 192.168.45.x to 192.168.44.x | 08:07 |
| persia | Assuming you have a single-gateway network, *other* hosts on 192.168.45.0/24 (assuming that's your network size) should use 192.168.45.x as a gateway. | 08:07 |
| persia | Can we use numbers? Your use of "x" confuses me. Let's call them 17 and 23. | 08:07 |
| persia | So 192.168.44.17 uses 192.168.44.2 as a gateway. | 08:07 |
| persia | 192.168.44.17 is the same host as 192.168.45.23 | 08:08 |
| persia | packets originating from that host should be routed to the internet through 192.168.44.2 | 08:08 |
| swift | it is? | 08:08 |
| persia | 192.168.45.19 is a compeltely different host. | 08:08 |
| persia | It uses 192.168.45.23 as a gateway. | 08:09 |
| persia | If you want the dual-homed host to *not* route IPv4 packets, fiddle with /proc/sys/net/ipv4/ip_forward | 08:09 |
| swift | yes.. I dont want the packets sent via 192.168.45.x to go through 192.168.44.x | 08:09 |
| swift | how do i guarantee that? | 08:09 |
| persia | So you want 192.168.45.19 to not have access to the internet? | 08:10 |
| alex_joni | echo "0" > /proc/sys/net/ipv4/ip_forward | 08:10 |
| swift | persia... say I ping another machine on the 192.168.45.0/24 network | 08:10 |
| persia | from where? | 08:10 |
| swift | it should go via 192.168.45.x on this machine | 08:10 |
| * persia waits for an example that *doesn't* use "x" to continue responding | 08:11 | |
| persia | (fake numbers are fine: I don't need the specifics, but I need something to usefully reference to distinguish the dual-homed host from other hosts) | 08:12 |
| alex_joni | swift: you only have one default gw that reaches 0.0.0.0 (e.g. internet) | 08:14 |
| persia | (usually) | 08:14 |
| alex_joni | for 192.168.45.0/24 you want a route that points to 192.168.45.0 | 08:14 |
| alex_joni | persia: right, for his case | 08:14 |
| persia | right. | 08:14 |
| alex_joni | so you should have: | 08:15 |
| alex_joni | Destination Gateway mask Iface | 08:15 |
| alex_joni | 192.168.44.0 * eth0 | 08:15 |
| alex_joni | 192.168.45.0 * eth1 | 08:15 |
| alex_joni | default 192.168.44.2 eth0 | 08:15 |
| alex_joni | forgot the mask in there.. | 08:16 |
| alex_joni | 255.255.255.0 for the first 2, and 0.0.0.0 for the last one | 08:16 |
| swift | if 192.168.45.10 is pinging 192.168.45.11(other machine)... it should use the 192.16845.x gateway... | 08:23 |
| _ruben | no, it shouldnt (need to) use a gateway at all | 08:24 |
| persia | Yes, and it ought do it automatically (assuming 192.168.45.10 is the dual-homed machine) | 08:24 |
| swift | yes.. | 08:25 |
| swift | but on the server the route shows it uses 192.168.44.2 as the gateway | 08:25 |
| persia | That's correct. | 08:25 |
| persia | So "gateway" has a couple potential interpretations. The simplest one is that it's the route used to get to 0.0.0.0 | 08:25 |
| _ruben | "the gateway" is kind of ambiguous (sp?) | 08:25 |
| binBASH | Hi | 08:25 |
| persia | swift: So I wouldn't worry much about 192.168.45.10: that machine sounds like it can get everywhere in all the correct ways. Check from 192.168.45.11 to verify it can ping 192.168.45.10 and can't ping 192.168.44.2 | 08:27 |
| persia | (where 192.168.45.11 should use 192.168.45.10 for 0.0.0.0) | 08:27 |
| swift | how can i use that interface? | 08:28 |
| swift | forgot ping syntax | 08:28 |
| swift | :D | 08:28 |
| persia | On 192.168.45.11 run `ping 192.168.44.2` | 08:29 |
| swift | persia.. I can see that it can ping a .44 IP on another machine!! | 08:31 |
| swift | i do 'ping -I eth1 192.168.44.13' | 08:31 |
| swift | it pings | 08:31 |
| persia | From 192.168.45.11? | 08:31 |
| swift | where eth1 is 192.168.45.11 | 08:31 |
| swift | yes | 08:31 |
| persia | OK. So, is this what you want? | 08:31 |
| swift | will I have to put a route there | 08:32 |
| persia | No, if you can ping, it already works. | 08:32 |
| swift | no... I shouldn't be able to communicate with 192.168.44.x network | 08:32 |
| swift | from 192.168.45.11... I should only be able to ping another 192.168.45.x IP | 08:32 |
| persia | OK, in that case, you want to disable IPv4 forwarding on 192.168.45.10. | 08:32 |
| * _ruben votes for a graphical network diagram and clear description of what should and should not be possible/allowed ;) | 08:33 | |
| persia | *OR* remove the route to 0.0.0.0 through 192.168.45.10 on 192.168.45.11. For extra points, do both. | 08:33 |
| stephank | The topic's tinyurl appears to be broken. | 08:37 |
| _ruben | seems to be cut off indeed | 08:38 |
| _ruben | adding "er" to it does work | 08:38 |
| stephank | Oh! I thought it was intentionally missing. :) | 08:39 |
| stephank | But is that really the intended page? It redirects to a blog post, rather than “doc and resources”. | 08:39 |
| _ruben | wouldnt know ;) | 08:40 |
| RoyK | hi all. any idea why /etc/postfix/main.cf seems to be missing in lucid? | 09:24 |
| persia | RoyK: I have it on a relatively fresh install for which I've done only minimal postfix configuration (setting up satellite system in the prompts). Did you maybe say "no configuration" on initial setup? | 09:25 |
| RoyK | I didn't install the box, but perhaps the guy who did so chose that - wouldn't surprise me | 09:26 |
| persia | RoyK: I just tested a fresh lucid postfix install in a chroot, and indeed, if one selects "No Configuration", no main.cf is produced (which seems like the correct behaviour to me). | 09:26 |
| persia | Try `dpkg-reconfigure postifx` | 09:27 |
| RoyK | persia: that was indeed the problem - the guy had just chosen "no config" because he didn't understand what to do and didn't bother to read those five lines | 09:30 |
| persia | RoyK: That's actually probably the safest thing that could have been done, and it makes it easy to fix (with dpkg-reconfigure). | 09:32 |
| RoyK | yeah | 09:34 |
| raphink | hi there | 10:04 |
| raphink | plymouth seems to be blocking startup on lucid | 10:04 |
| raphink | anybody witnessed that? | 10:04 |
| persia | raphink: That's almost never actually plymouth. What's the specific issue? | 10:18 |
| raphink | when we reboot the server | 10:18 |
| raphink | almost no service is started | 10:18 |
| raphink | except ntp | 10:18 |
| raphink | syslog-ng is not started for example | 10:18 |
| raphink | and plymouthd keeps running | 10:18 |
| raphink | after I log in ssh | 10:18 |
| raphink | deactivating plymouth in /etc/init seemed to help with quite a few services | 10:19 |
| persia | Odd. All plymouth should be doing is IO multiplexing (and maybe some screen draws). | 10:19 |
| persia | Are you sure something else isn't hanging on IO, and it gets a bit further when it crashes on an IO failure when plymouth isn't around? | 10:20 |
| raphink | I've tried to put bootchart | 10:22 |
| raphink | to log what's happening during the boot | 10:22 |
| raphink | but it doesn't generate the png | 10:22 |
| raphink | it crashes quite badly | 10:22 |
| persia | Can you get the bootlog info, and then generate the graphics post-boot? | 10:24 |
| persia | Also, have you filed a bug? This doens7t happen for my serers, and it sounds potentially serious. | 10:25 |
| raphink | empathy crashed ;-) | 10:27 |
| persia | Heh. You're having great luck today :) | 10:28 |
| persia | Have you filed a bug yet? | 10:28 |
| persia | (about the boot) | 10:28 |
| raphink | no, I've been searching through the bugs on LP | 10:29 |
| raphink | ideally, we'd get rid of plymouth | 10:29 |
| raphink | but there's a bug that is marked as invalid about too strict deps on plymouth | 10:29 |
| persia | No, plymouth is essential. Nothing else handles boot-time IO. Using text mode, it doesn't even do much GUI stuff. | 10:30 |
| persia | I'd recommend filing a bug against upstart: it shouldn't half-start a system without giving you useful output. | 10:31 |
| raphink | well currently we haven't checked the console output | 10:31 |
| raphink | we'll go to the machine room this afternoon hopefully to have a look at it | 10:32 |
| persia | That's worth checking: I suspect that it will give you a better target for the bug report. | 10:34 |
| raphink | ok | 10:35 |
| raphink | thanks for your time | 10:35 |
| === diehaai is now known as thefish | ||
| bronto2 | what would one enable on apache to get .htaccess stuff working? i just need to pass protect a certain folder | 11:17 |
| _ruben | AllowOverrides | 11:18 |
| _ruben | and you dont need .htaccess for that per se, can be done in (global) config just as well | 11:19 |
| bronto2 | _ruben, yes i see, seems everything is configured, but still not working | 11:23 |
| _ruben | not working? tell it to get a job then | 11:29 |
| binBASH | _ruben: /etc/glusterfs/store01-tcp.vol 8.1T 116M 8.1T 1% /mnt/storage | 11:29 |
| binBASH | working with the lucid packages ;) | 11:29 |
| _ruben | binBASH: nice, i should run some tests with it as well, might have some usecases for it | 11:30 |
| binBASH | _ruben: Well I only run it with 100 Mbit though | 11:30 |
| binBASH | provider doesn't have gbit :p | 11:30 |
| binBASH | so only 5.8 MB /sec for writes | 11:31 |
| binBASH | but I don't need fast storage | 11:31 |
| binBASH | can live with that. | 11:31 |
| bronto2 | _ruben, i have some other users, that cannot access global conf, so i'd like this to work via htaccess | 11:37 |
| bronto2 | anything wrong with this http://pastebin.com/fq9ZVSwP ? | 11:37 |
| === libertiy is now known as libertiy|away | ||
| _ruben | bronto2: looks good, tho it being only a partial config, cant say anything definitive about it | 11:40 |
| bronto2 | hmm, there is no AllowOverrides directive in apache2.conf | 11:41 |
| bronto2 | but should defaults to ALL right? | 11:41 |
| _ruben | a sane config sets the default to None | 11:43 |
| === libertiy|away is now known as libertiy | ||
| smoser | mdz, fyi, i found a bug that had us not un-publishing our daily images, which caused the ramp up | 13:20 |
| smoser | in ec2 images as reported by cloudmarket | 13:20 |
| mdz | smoser, nice work | 13:21 |
| mdz | might save us a few pennies on our S3 bill too ;-) | 13:21 |
| lau | hello, runing hardy + samba 3.0.28 I have an error log like | 13:48 |
| lau | create_builtin_users: Failed to create Users | 13:48 |
| lau | any idea to fix this please ? | 13:48 |
| ttx | smoser: any reason why "Ubuntu Server EC2 EBS (Europe) amd64 (ami-8b705aff)" has no registered results so far ? | 13:50 |
| smoser | humans suck | 13:50 |
| smoser | i tested, just didn't record | 13:50 |
| smoser | ttx, fixed | 13:51 |
| ttx | smoser: thanks ! | 13:51 |
| ttx | smoser: I hit the "no ssh connection" issue on amd64 UEC images quite a few times, I confirm it looks like a timing issue, the slower i386 image didn't hit it in my tests | 13:52 |
| smoser | ttx, did you get console logs ? | 13:53 |
| smoser | because there are 2 possible issues | 13:53 |
| smoser | and one i have much less understanding on | 13:53 |
| ttx | smoser: yes, I have it, let me pastebin it | 13:54 |
| ttx | smoser: http://pastebin.ubuntu.com/419841/ | 13:55 |
| === ogra_ is now known as ogra | ||
| ttx | it's the test with your test userdatafile | 13:55 |
| smoser | is that system u p? | 13:55 |
| ttx | smoser: shows hello world alright | 13:55 |
| ttx | smoser: no longer | 13:55 |
| ttx | i can reproduce it if need be though | 13:56 |
| smoser | thats bug https://bugs.launchpad.net/ubuntu/+source/cloud-init/+bug/565018 | 13:56 |
| uvirtbot | Launchpad bug 565018 in cloud-init "instance is not reachable via ssh" [Undecided,New] | 13:56 |
| smoser | and i have been extremely unsuccessful in reproducing it with some debug on | 13:56 |
| smoser | i've run probably 2000 instances trying to | 13:56 |
| ttx | smoser: it's also a metadata service issue ? | 13:57 |
| smoser | no. i dont think so. | 13:57 |
| smoser | i have no clue on this. | 13:57 |
| ttx | try 19/20 ? | 13:57 |
| smoser | wait, where ? | 13:57 |
| ttx | smoser: in your bug comments | 13:58 |
| smoser | (note, the 19/20 stuff is just bad, it zero based counting) | 13:58 |
| smoser | ttx, the comment... the 25 cases that have that in their logs are *not* your bug | 13:59 |
| smoser | that you're sseeing, and that i can't reproduce easily | 13:59 |
| smoser | i have some info i'd like to get if it reproduces | 13:59 |
| smoser | 1.) want to collect successfully the error messages from ssh when you connect | 13:59 |
| ttx | I can reproduce it if you need | 14:00 |
| smoser | 2.) want to try ssh-keyscan and get those errors | 14:00 |
| smoser | 3.) ideally turn debug on in upstart, but that seems to reduce likelyhood of error | 14:00 |
| smoser | ttx, for 1 and 2 if you can get a recreate, please collect | 14:00 |
| ttx | on my way | 14:01 |
| smoser | for 3 if you can install the cloud-init from https://bugs.launchpad.net/ubuntu/+source/cloud-init/+bug/566792 | 14:01 |
| uvirtbot | Launchpad bug 566792 in eucalyptus "UEC guests sometimes fail on consuming user data (metadata service isn't ready)" [High,Confirmed] | 14:01 |
| smoser | when you dpkg -i that, it will tell you how to enable upstart debug | 14:01 |
| smoser | the problem with upstart debug is that due to bug 566793 you can't get all the console data if you use upstart debug | 14:02 |
| uvirtbot | Launchpad bug 566793 in eucalyptus "euca-get-console-output gives first 64k of output, not most recent" [Low,New] https://launchpad.net/bugs/566793 | 14:02 |
| smoser | its too verbose | 14:02 |
| ttx | smoser: hm, can't reproduce it now | 14:05 |
| smoser | maybe you need to change the way you're sitting in your seat | 14:05 |
| smoser | :-( | 14:06 |
| uvirtbot | New bug: #527933 in bacula (main) "wuiso" [Low,Invalid] https://launchpad.net/bugs/527933 | 14:06 |
| smoser | i would really like to get debug on this, because I can't come up with any idea as to why its not working. it seems to me that the cloud-init-cfg stuff is just not getting started (ie, upstart jobs not running) | 14:06 |
| * ttx grumbles | 14:07 | |
| * ttx scraps his topology3 setup that now works like a clockwork and tries a topology 2 | 14:09 | |
| smoser | ttx, for this i really dont think its topology based at all | 14:10 |
| smoser | i think its racey inside the image itself | 14:10 |
| ttx | smoser: sure, but I need to cover the missing ISo tests more than I need to chase heisenbugs | 14:12 |
| smoser | because cloud-init ran, and sshd ran (runs on 'filesystem' event, same as the cloud-init-cfg events) | 14:12 |
| ttx | one never knows, it might reappear and be time-since-registration-sensitive | 14:12 |
| smoser | or due to having '.tar' regex in the manifest name :) | 14:12 |
| ttx | smoser: do I detect some sarcasm ? :) | 14:12 |
| ttx | how can you tell, looking at the console logs, the difference between bug 565018 and bug 566792 ? | 14:15 |
| uvirtbot | Launchpad bug 565018 in cloud-init "instance is not reachable via ssh" [High,Confirmed] https://launchpad.net/bugs/565018 | 14:15 |
| ttx | smoser: ^ | 14:15 |
| uvirtbot | Launchpad bug 566792 in eucalyptus "UEC guests sometimes fail on consuming user data (metadata service isn't ready)" [High,Confirmed] https://launchpad.net/bugs/566792 | 14:15 |
| smoser | ttx, look at http://launchpadlibrarian.net/44832884/guest-error.log and http://launchpadlibrarian.net/44479315/console.fail.1.txt | 14:16 |
| smoser | firs is the "fail on consuming" | 14:16 |
| smoser | see, it falls all over the place , all the cloud-init-cfg stuff runs, but fails because its got no metadata | 14:17 |
| smoser | in the second, the stuff either runs quietly to success or not at all | 14:17 |
| ttx | smoser: it's that verbose even with the default image ? | 14:17 |
| smoser | there are no python trace | 14:17 |
| smoser | yeah... the python throws error | 14:17 |
| ttx | ok, then I definitely have been hitting 565018 | 14:18 |
| ttx | I don't remember ever seeing those python stacktraces in console logs | 14:18 |
| * ttx fixes iso test results accordingly | 14:18 | |
| smoser | well, if/when you do, please get 'ssh -v instance-ip 2>err.txt 1>out.txt' and 'ssh-keyscan instance-ip 2>err.keyscan 1>out.keyscan' please | 14:19 |
| smoser | ttx, well, you shoudln't see them :) | 14:19 |
| ttx | excellent. | 14:19 |
| baffle | kirkland/nijaba: RH has produced the library "netcf" [https://fedorahosted.org/netcf/] for configuring/polling network configuration. This library is (in theory) distribution-agnostic, but lacks a backend for Ubuntu/Debian. The library is used by libvirt to support advanced (remote) network configuration/usage, and virt-manager now expose this as well. There is also talk about using it in NetworkManager [https://fedoraproject.org/wiki/Features/Shared_N | 14:20 |
| baffle | "Someone" should write a new backend for Netcf so that libvirt in Ubuntu can use these new features. It is written in C, and will need a developer that also is knowledgeable in network configuration "the Debian way". | 14:20 |
| baffle | I propose that this is written up for discussion at UDS. I do not think the community will develop this, due to it being so tightly coupled with the distribution. | 14:20 |
| baffle | (Okay, this is possibly ML material, but I don't really hang around in them much.. :-)) | 14:20 |
| ttx | zul: could you cover the amd64/printserver and amd64/postgresql ISO tests ? | 14:21 |
| zul | ttx: yep | 14:21 |
| diago | in fstab for lvm do you use /dev/lvm-raid/shares or the /dev/mapper device? | 14:31 |
| kirkland | baffle: thanks, answered in #ubuntu-virt, though I will copy here for everyone else's benefit | 14:34 |
| kirkland | <kirkland> nijaba: baffle: hi guys, I am familiar with netcf, i'm tracking its progress | 14:34 |
| kirkland | <kirkland> nijaba: baffle: I'd very much like to see it packaged and used in Ubuntu in Maverick | 14:34 |
| kirkland | <kirkland> nijaba: baffle: that's not a commitment, though | 14:34 |
| kirkland | <kirkland> nijaba: baffle: but I will put together a spec | 14:34 |
| baffle | kirkland: Wunderbar (from #ubuntu-virt) | 14:34 |
| kirkland | baffle: ;-) | 14:35 |
| kirkland | baffle: what's your LP id? | 14:35 |
| kirkland | baffle: i'll subscribe you to the blueprint (once I get around to writing it) | 14:35 |
| baffle | kirkland: Hmm. Either dag-stenstad, or dag.stenstad@ventelohosting.no or baffle. :-) Not sure wich is my ID, really. OpenID is https://launchpad.net/~dag-stenstad | 14:37 |
| kirkland | baffle: cheers | 14:37 |
| baffle | kirkland: I'm so confused! Wich is my ID? :) | 14:40 |
| kirkland | baffle: s/^.*~// | 14:40 |
| kirkland | baffle: dag-stenstad | 14:40 |
| ummagumma | hi everybody! does anyone knows a channel for wxpython without the need to get a registry? | 14:46 |
| ttx | smoser: meeting time ? | 15:01 |
| ttx | (nm) | 15:01 |
| osmosis | kirkland, where do we register our test results? also, sounds like UEC testings is high priority for lucid release. | 15:11 |
| kirkland | osmosis: http://iso.qa.ubuntu.com/ | 15:12 |
| uvirtbot | New bug: #494141 in samba (main) "CUPS starts after SAMBA; printers are not available" [Undecided,New] https://launchpad.net/bugs/494141 | 15:16 |
| smoser | kirkland, the oops... its the emulated scsi device | 15:26 |
| kirkland | smoser: oh, is it? | 15:26 |
| kirkland | smoser: well that's not a kernel thing then | 15:26 |
| smoser | yeah. i duped hggdh's to my bug that i'd previously opened on that. | 15:27 |
| smoser | searching for numbers | 15:27 |
| smoser | bug 564924 is dupe of bug 546458 | 15:27 |
| uvirtbot | Launchpad bug 564924 in linux "UEC guest sometimes gets kernel OOPS (dup-of: 546458)" [Medium,Confirmed] https://launchpad.net/bugs/564924 | 15:27 |
| uvirtbot | Launchpad bug 546458 in qemu-kvm "kernel NULL pointer in -virtual (-server) kernel" [Low,Confirmed] https://launchpad.net/bugs/546458 | 15:27 |
| eagles05138785 | hummm what am i looking for | 16:12 |
| eagles05138785 | !lucid | 16:12 |
| ubottu | Lucid Lynx is the codename for Ubuntu 10.04, due April 29th, 2010 - Lucid is NOT released and is NOT stable - Discussion and support only in #ubuntu+1 - Development Schedule: https://wiki.ubuntu.com/LucidReleaseSchedule | 16:12 |
| eagles05138785 | not what i want either | 16:13 |
| eagles05138785 | anyone have a link to where i can get the beta 2 of ubuntu server | 16:13 |
| jdstrand | kirkland, smoser, ttx: fyi, I added an fd leak check to test-libvirt.py in QRT. it only tests start/destroy, but more can be added if needed | 16:17 |
| jdstrand | that will hopefully help moving forward | 16:17 |
| kirkland | jdstrand: neat :-) | 16:17 |
| cloakable | 8 days :D | 16:17 |
| Jeeves_ | RC tomorrow | 16:18 |
| guntbert | eagles05138785: http://cdimage.ubuntu.com/ubuntu-server/daily/current/ | 16:18 |
| eagles05138785 | ty guntbert | 16:18 |
| guntbert | eagles05138785: you're welcome :-) | 16:19 |
| eagles05138785 | Jeeves_: would rather get it now and have something to install with then wait till release when everyone is fighting for a piece of lucid | 16:19 |
| eagles05138785 | fighting for a piece of the lynx haha | 16:19 |
| cloakable | heh | 16:19 |
| * cloakable will wait a bit | 16:19 | |
| shrini | team: need help on nfs client | 16:20 |
| eagles05138785 | cloakable: regardless of it being the beta u install with or the rc you still get the same updates | 16:20 |
| eagles05138785 | thanks again guntbert :) | 16:20 |
| shrini | My share is disconnected intermettenlt | 16:20 |
| shrini | y | 16:20 |
| shrini | need help on finding the casue | 16:20 |
| cloakable | eagles05138785: Yeah, but I'd rather wait for stable, then the first round of updates to make sure it's stable :P | 16:21 |
| eagles05138785 | everyone to their own so i say hehe | 16:21 |
| Jeeves_ | eagles05138785: I'm allready running it for about 5 months | 16:21 |
| eagles05138785 | ya been on it since early alphas | 16:21 |
| eagles05138785 | since late alphas on my server early alphas on vm on my desktop running kubuntu | 16:21 |
| eagles05138785 | would rather do a clean install on my server then upgrading so i dont have any reminants of karmic around hehe | 16:21 |
| ajsie | what have they improved in ubuntu server 10.4? | 17:02 |
| hggdh | kirkland: really, eucalyptus is not doing garbage collection. Memory usage at the CC and CLC is still high, after 5+ hours of idling | 17:07 |
| jdstrand | kirkland: I noticed you said that you wanted to fix bug #562266 | 17:07 |
| uvirtbot | Launchpad bug 562266 in libvirt "Make virt-pki-validate usable on Ubuntu" [Medium,Triaged] https://launchpad.net/bugs/562266 | 17:07 |
| jdstrand | kirkland: fyi: ubuntu26 is already in the queue and it might be nice to fix bug #565380 in the process | 17:09 |
| kirkland | jdstrand: what do you think? | 17:09 |
| uvirtbot | Launchpad bug 565380 in libvirt "postinst fails if missing user is in the admin group" [Low,Triaged] https://launchpad.net/bugs/565380 | 17:09 |
| jdstrand | I'm fine with fixing these things before release, cause it is easier than SRU, taking into account risk of regression, etc , etc | 17:09 |
| jdstrand | kirkland: 565380 is super low risk (we add an '|| true') | 17:10 |
| kirkland | jdstrand: gotcha | 17:11 |
| kirkland | jdstrand: sure, assign that to me, i'll do an upload that fixes those two | 17:11 |
| kirkland | jdstrand: i'm doing an interview right nw | 17:11 |
| jdstrand | kirkland: mostly I just wanted to make sure you grabbed ubuntu26 before doing anything :) | 17:11 |
| kirkland | jdstrand: of course ;-) | 17:11 |
| jdstrand | kirkland: k | 17:11 |
| kirkland | jdstrand: you're welcome to snag those bugs if you like, too :-) | 17:11 |
| jdstrand | well, I meant to do the 565380, but forgot | 17:12 |
| jdstrand | as for the other, I don't know much about it tbh | 17:12 |
| maddhat | Hey everyone, trying to get ubuntu server karmic netboot working but it says it cant find the ethernet driver. any way to manually add it to the boot img? | 17:29 |
| uvirtbot | New bug: #567623 in postfix (main) "package postfix 2.6.5-3 failed to install/upgrade: subprocess installed post-installation script returned error exit status 75" [Undecided,New] https://launchpad.net/bugs/567623 | 17:44 |
| === Guest123456 is now known as Guest84312 | ||
| sabator | Where is the best tutorial for gdm over ssh ? | 17:49 |
| RoyK | sabator: gdm over ssh? | 18:06 |
| RoyK | or just X over ssh? | 18:06 |
| bluethundr_ | I am attempting to install courierauthlib from source because i would like some finer grain control over the options... configure runs ok, but afterwards when I try to make the app it craps out, and I could use some help interpreting the output.. | 18:08 |
| bluethundr_ | http://pastebin.com/avPiV9w4 | 18:08 |
| RoyK | perhaps handle-exceptions? | 18:10 |
| Guest95767 | I am trying to install Ubuntu-or-Debian server in a Xen Guest, on a non-Debian/Ubuntu Xen host. How/where do I tell the Debian Xen guest's installer to use a preseed file on an nfs volume? Some parameter at 'xm create ... install-extra="..."' &/or the Xen cfg's 'extra = ...' clause. | 18:10 |
| RoyK | -f | 18:10 |
| Guest95767 | So far, I've failed to get the guest to see the nfs mount. It *does* launchg the manual/interactive installer,though ... | 18:10 |
| RoyK | bluethundr_: try to remove -fhandle-exceptions | 18:11 |
| === luis__lopez is now known as luis_lopez | ||
| bluethundr_ | RoyK: what is the best way to accomplish this? (i.e. remove -fhandle-exceptions) | 18:16 |
| RoyK | bluethundr_: in /home/bluethundr/courier-authlib-0.63.0/gdbmobj/Makefile I guess | 18:17 |
| === dendrobates is now known as dendro-afk | ||
| bluethundr_ | RoyK: that did it...thanks | 18:38 |
| smoser | kirkland, or ttx, how do i manually register an NC ? | 18:51 |
| smoser | it seems it didn't "just work". that might be because i re-installed it, but on the re-instal it didn't get registerd | 18:51 |
| smoser | never mind. it did. user error. | 18:53 |
| === libertiy is now known as libertiy|away | ||
| electro_ | Is there a channel for EUC ? | 19:11 |
| RoyK | electro_: #ubuntu-could, but this one works as well | 19:15 |
| electro_ | RoyK: are you running an EUC cloud? | 19:30 |
| RoyK | electro_: no | 19:31 |
| RoyK | electro_: I was looking into it, but it seems hard to make it truely redundant, so I guess we'll find another solution | 19:31 |
| _Trullo | I installed ubuntu server, did apt-get install xinit , now when I startx a error message, xrdb command not found, X not merged | 19:32 |
| RoyK | _Trullo: if you want X on your server, I'd say Ubuntu Desktop is a better way to start | 19:32 |
| RoyK | there aren't really much difference between ubuntu server and desktop, just a few kernel changes, and you can install ubuntu server kernel later | 19:33 |
| _Trullo | I just want X to work | 19:33 |
| RoyK | gui on the server? | 19:33 |
| electro_ | RoyK: what problems does it have with redundancy? | 19:33 |
| _Trullo | probably won't be there anyway | 19:33 |
| electro_ | RoyK: Im interested to see how it handles a master failure | 19:33 |
| RoyK | electro_: there really isn't a good way to make the front-end redundant | 19:33 |
| electro_ | yeah | 19:33 |
| Hypnoz | Upgrading a server from 7.10 to 8.04, I copied the 8.04 sources.list, then did apt-get update && apt-get upgrade | 19:34 |
| Hypnoz | however a lot of packges were "kept back" | 19:35 |
| RoyK | Hypnoz: the right way is 'do-release-upgrade' | 19:35 |
| RoyK | or is that just in newer distros? | 19:35 |
| persia | That's in newer things. | 19:35 |
| RoyK | ok | 19:35 |
| Hypnoz | apt-get dist-upgrade? | 19:35 |
| persia | But "newer" in this case appears to include 7.10 :) | 19:35 |
| RoyK | hehe | 19:36 |
| persia | Hypnoz: More of a concern is that do-release-upgrade appears to upgade you to the latest release (9.10 in this case). | 19:37 |
| electro_ | RoyK: is there anything better you are testing? | 19:37 |
| Hypnoz | persia: exactly, I didn't want to upgrade to the newest ubuntu release, just 8.04 LTS | 19:37 |
| RoyK | electro_: we might be using xen or perhaps paying a lot for vmware | 19:37 |
| persia | Hypnoz: apt-get dist-upgrade is probably safest (but I may be wrong). Once you have 8.04, you want to do do-release-upgrade to get to 10.04 (post release) | 19:38 |
| Hypnoz | so I put that sources.list in and did apt-get upgrade | 19:38 |
| Hypnoz | I don't want 10.04 | 19:38 |
| persia | apt-get upgrade is too careful. | 19:38 |
| Hypnoz | i want 8.04 LTS | 19:38 |
| RoyK | apt-get dist-upgrade is a start | 19:38 |
| Hypnoz | will dist-upgrade push me past 8.04? | 19:38 |
| persia | Hypnoz: Right. Not now. Later. | 19:38 |
| RoyK | do-release-upgrade takes you to the next (lts) release | 19:39 |
| persia | Hypnoz: Use `apt-get dist-upgrade` for now. | 19:39 |
| RoyK | Hypnoz: no | 19:39 |
| RoyK | Hypnoz: it'll take you to the next lts | 19:39 |
| RoyK | the last lts is 8.04 | 19:39 |
| Hypnoz | ok | 19:39 |
| persia | Hypnoz: Sometime between now and the release of 12.04, you'll find your life simpler if you later run do-release-upgrade to get to 10.04 LTS. You may want to wait up to two years for this. | 19:39 |
| RoyK | but roll back your sources.list first | 19:39 |
| persia | RoyK: Are you sure? I thought do-release-upgrade went to LTS only for LTS installs, so a 7.10 install would try repeated upgrades to get to 9.10. | 19:40 |
| electro_ | RoyK: Im thinking you can do some sort of Active / Passive failover with the front-end | 19:40 |
| RoyK | active | 19:41 |
| _ruben | afaik, do-release-upgrade get you the next version, which by default is the next lts, but only for lts releases :) | 19:41 |
| RoyK | persia: see /etc/update-manager/release-upgrades | 19:41 |
| electro_ | RoyK: so is EUC pretty much a VM infrastructure? | 19:41 |
| persia | RoyK: Ah, right. | 19:41 |
| persia | Hypnoz: Check that file and make *sure* it has "Prompt=lts" | 19:41 |
| RoyK | electro_: it is, but it contains a single point of failure, the front-end | 19:41 |
| persia | Hypnoz: If that has prompt=lts, do-release-upgrade should bring you to 8.04 LTS (as long as you run it before 10.04 LTS releases) | 19:42 |
| Hypnoz | persia: in sources.list? | 19:42 |
| RoyK | electro_: so unless you use drbd+heartbeat or something for that, all your VMs will go down if that box dies | 19:42 |
| persia | Hypnoz: In /etc/update-manager/release-upgrades | 19:43 |
| kirkland | ttx: fyi, i walked smoser through his node registration issue privately | 19:44 |
| _ruben | do-release-upgrade wont allow you to skip releases (except for lts to next lts), so 7.10 should never get upgraded to 10.04 directly, or 9.10 for that matter | 19:44 |
| Hypnoz | persia: there is no /etc/update-manager.. | 19:45 |
| persia | Ah, then 7.10 doesn't have that. | 19:45 |
| Pici | do-release-upgrade will also prompt you before it does anything. | 19:45 |
| persia | Yeah, modify sources.list and run `apt-get dist-upgrade` then. | 19:45 |
| persia | Pici: But does it let you decide to only update to 8.04 if you want? | 19:46 |
| ttx | kirkland: ack | 19:46 |
| electro_ | RoyK: do you have any documentation stating this redundancy issue? I cant seem to google anything | 19:46 |
| Hypnoz | wish apt-get dist-upgrade told me what dist it was going to upgrade to | 19:46 |
| RoyK | electro_: it's pretty well known | 19:46 |
| _ruben | also: https://help.ubuntu.com/community/HardyUpgrades#Network%20Upgrade%20from%207.10%20for%20Ubuntu%20Servers%20%28Recommended%29 | 19:46 |
| _ruben | basically: install update manager, run update manager ;) | 19:47 |
| Hypnoz | _ruben: now that i've replaced sources.list with 8.04's, and did apt-get upgrade, do-release-upgrade shows "No new release found" | 19:49 |
| Hypnoz | _ruben: I think I would have had to do that first | 19:49 |
| Hypnoz | and cat /etc/issue shows 8.04.4 LTS so the system is upgraded, it just held back a lot of packages for some reason | 19:49 |
| RoyK | lsb_release -r | 19:50 |
| _ruben | Hypnoz: editing your sources.list to a newer version + running upgrade instead of dist-upgrade is a recipe for disaster | 19:50 |
| persia | Hypnoz: I think you need to dist-upgrade to get out of this state. `apt-get upgrade` is never enough to upgrade release-to-release. | 19:50 |
| _ruben | Hypnoz: revert your sources.list and hope do-release-upgrade can fix the mess | 19:50 |
| Hypnoz | RoyK: "Release:8.04" | 19:50 |
| persia | I don't think reversion is an option at this point: too much has been upgraded. | 19:51 |
| _ruben | Hypnoz: or if feeling brave, run apt-get dist-upgrade and hope it wont break any further | 19:51 |
| _ruben | when it does, you get to keep both/all pieces, though | 19:51 |
| RoyK | Hypnoz: rebooted yet? | 19:51 |
| Hypnoz | no i thought about that and doing another apt-get upgrade | 19:51 |
| persia | Don't reboot! | 19:51 |
| persia | And apt-get upgrade isn't going to help at this point. | 19:52 |
| persia | It will not handle the transitions. | 19:52 |
| persia | You need dist-upgrade for apt to do anything. | 19:52 |
| _ruben | apt-get upgrade was never meant to change versions, that's what dist-upgrade was originally made for | 19:52 |
| persia | do-release-upgrade might have worked, but ... | 19:52 |
| _Trullo | is ctrl-alt-del disabled in ubuntu-server? :) | 19:52 |
| Hypnoz | no | 19:53 |
| persia | Just check the output when apt tells you want it plans to do very carefully. | 19:53 |
| RoyK | don't think so | 19:53 |
| _ruben | do-release-upgrade basically does a dist-upgrade, but has some extra tricks up its sleeve to handle known upgrade problems | 19:53 |
| _ruben | _Trullo: not by default, but can be done | 19:53 |
| RoyK | _Trullo: it was changed to ctrl+alt+shift+esc+backspace+f+u+v+m | 19:53 |
| Hypnoz | persia: result of apt-get dist-upgrade http://dpaste.de/FRw9/ | 19:53 |
| RoyK | Hypnoz: Y | 19:54 |
| _Trullo | hehe | 19:54 |
| _ruben | Hypnoz: looks sane to me | 19:54 |
| Hypnoz | it doesn't tell me what release its going to put me at | 19:54 |
| persia | Hypnoz: Looks reasonably safe to me: nothing is being removed. | 19:54 |
| _ruben | (considering the circumstances) | 19:54 |
| RoyK | Hypnoz: it doesn't | 19:54 |
| Hypnoz | it could keep me at 8.04 or send me to 9.10 | 19:55 |
| RoyK | 8.04 afaik | 19:55 |
| _ruben | Hypnoz: dist-upgrade isnt gonna change releases for you, ever, not without you editing the sources.list | 19:55 |
| RoyK | _ruben: didn't he change that? | 19:55 |
| Hypnoz | yeah i put sources.list to 8.04 | 19:55 |
| _ruben | RoyK: to 8.04 yes, so that's where he'll stick | 19:55 |
| persia | Then dist-upgrade will take you to 8.04. | 19:55 |
| RoyK | yeah | 19:55 |
| persia | But next time you upgrade, please consider do-release-upgrade :) | 19:56 |
| _ruben | (in a non-supported way though) | 19:56 |
| persia | Right. | 19:56 |
| _ruben | +1 | 19:56 |
| persia | Well, we're kinda supporting it, but we offer no guarantee it won't break. | 19:56 |
| RoyK | looked a little small amount of packages to take him all the way to 8.04, though | 19:56 |
| _ruben | no guarantees for do-release-upgrade either afaik ;) | 19:56 |
| Hypnoz | RoyK: a lot were upgraded in apt-get upgrade | 19:57 |
| persia | RoyK: Remember, it'S post `apt-get upgrade`: this is just the leftover transition stuff. | 19:57 |
| RoyK | final solution: install Windoze! | 19:57 |
| persia | heh, no. | 19:57 |
| Hypnoz | RoyK: haha | 19:57 |
| _ruben | apt-get upgrade did the "safe" part .. dist-upgrade takes care of the more trickier parts | 19:57 |
| persia | No, it's not that smart. It just does the unsafe parts. | 19:57 |
| RoyK | apt-get upgrade takes you to the latest x.x.x, where dist-upgrade takes you to the latest x.x | 19:58 |
| persia | But for this specific set of packages, it seems to be only removing an obsolete library, and an obsolete way of identifying storage devices. | 19:58 |
| _ruben | RoyK: eh? | 19:58 |
| persia | RoyK: Hrm? No. | 19:58 |
| RoyK | as upgrade takes you to the latest 8.04.2 and dist-upgrade takes you from 8.04.2 to 8.04.3 | 19:58 |
| persia | RoyK: safe-upgrade will never remove anything. dist-upgrade will. | 19:58 |
| Hypnoz | seems like updating sources.list to 8.04 was correct, i should have just done dist-upgrade instead of upgrade | 19:58 |
| _ruben | Hypnoz: not really ;) | 19:59 |
| _ruben | install update-manager-core + running do-release-upgrade would've been correct ;) | 19:59 |
| qman__ | upgrade also won't install new packages, such as kernel updates, where dist-upgrade will | 19:59 |
| _ruben | or atleast the recommended approach | 19:59 |
| persia | Hypnoz: Ideally you would have run do-release-upgrade, but yeah, *if* you've already fiddled your sources.list and want to upgrade with apt, dist-upgrade is the correct command. | 19:59 |
| persia | qman__: Oh, new dependencies are blocked also? That makes sense. | 20:00 |
| _ruben | correct | 20:00 |
| _ruben | kernel upgrade without abi changes are taken care of by upgrade iirc | 20:01 |
| RoyK | apt-get install windowsxp | 20:01 |
| Hypnoz | sweet after dist-upgrade I'm at Ubuntu 8.04.4 LTS \n \l | 20:03 |
| Hypnoz | and apt-get upgrade shows "0 upgraded, 0 newly installed, 0 to remove and 0 not upgraded" | 20:03 |
| Hypnoz | hooray for me not being fired! | 20:03 |
| Hypnoz | ;) | 20:03 |
| _ruben | Hypnoz: you got fairly lucky though ;) | 20:04 |
| _ruben | apparently your set of installed packages isnt complex enough for dist-upgrade to fail ;) | 20:04 |
| _ruben | (dist-upgrade with altered sources.list that is) | 20:04 |
| Hypnoz | I guess, but dist-upgrade was made for this purpose | 20:05 |
| Hypnoz | to have it not work would be silly | 20:05 |
| _ruben | Hypnoz: its ancient and cant handle certain upgrades/packaging changes, that's why do-release-upgrade was invented, to work around those issues | 20:05 |
| Hypnoz | _ruben: sounds good i'll keep it in mind next time | 20:06 |
| persia | Hypnoz: You really should have an /etc/update-manager directory now. | 20:07 |
| persia | You're running 8.04, but you really want "Prompt=lts" in there if you want to be running 8.04 LTS. | 20:07 |
| persia | Just so the next upgrade is really to 10.04 LTS, and not to something else. | 20:08 |
| persia | (not that this is going to happen anytime soon) | 20:08 |
| Hypnoz | 10.04 LTS won't happen within the next 6 months? | 20:08 |
| jpds | Hypnoz: More like next week Thursday. | 20:09 |
| persia | The release is the 29th, but as you're just upgrading from 7.10 today, I'm guessing you won't plan on upgrade to 10.04 for a bit :) | 20:10 |
| mathiaz | Hypnoz: *upgrading* from 8.04 LTS to 10.04 LTS should be automatically enabled in July when 10.04.1 (the first point release) is released | 20:10 |
| mathiaz | Hypnoz: note that upgrading from 8.04 LTS is already possible (and suggested) | 20:11 |
| mathiaz | Hypnoz: the end user is not automatically offered the choice yet | 20:11 |
| _ruben | mathiaz: not quite following you here, what exactly will change at 10.04.1 ? | 20:11 |
| mathiaz | _ruben: when you open update-manager and you've configured your system to only upgrade from LTS to LTS, you'll only see a message stating that there is a new release available around 10.04.1 | 20:12 |
| _ruben | mathiaz: ahh ok | 20:13 |
| _ruben | sounds sane enough :) | 20:13 |
| mathiaz | _ruben: people running karmic will see a message that a new release is available as soon as 10.04 is available | 20:13 |
| _ruben | also sane enough :) | 20:13 |
| smoser | ok. google is fialing me. is there a way to preseed an apt-get upgrade ? i know that it is going to prompt for 2 questions. 'grub2/linux_cmdline' for example. | 20:13 |
| smoser | is there a way to feed my answer , like preseed install, but for apt or dpkg | 20:14 |
| mathiaz | smoser: debconf-set-selections | 20:14 |
| persia | smoser: Sure. You can preseed any debconf value. | 20:14 |
| smoser | thats what i wanted. debconf-set-selections. thank you mathiaz | 20:14 |
| smoser | thanks persia | 20:14 |
| mathiaz | smoser: that should work for most of the packages | 20:15 |
| mathiaz | smoser: some packages might reset the debconf database | 20:15 |
| persia | `echo "buildd shared/accepted-sun-dlj-v1-1 boolean true" | debconf-set-selections` is the trick we used to use in schroots to build Java stuff. | 20:15 |
| mathiaz | smoser: but that's a corner case | 20:15 |
| persia | mathiaz: Isn't that typically a bug? | 20:16 |
| mathiaz | persia: yeah - I agree | 20:16 |
| mathiaz | persia: I haven't seen this though | 20:16 |
| mathiaz | persia: It would be possible - but I agree it should be a bug | 20:16 |
| persia | mathiaz: I know a package that does it, but it's not possible to install an Ubuntu system without that package, and that package has a well-documented way to preseed the values it doesn't reset. | 20:17 |
| persia | But I'd *really* not like to see that used for packages that didn't meet that set of criteria :) | 20:17 |
| mathiaz | persia: which package are you refering to? | 20:17 |
| persia | console-setup | 20:17 |
| persia | It rests *part* of its debconf stuff while doing discovery against it's (documented) preseeding. | 20:18 |
| ziesemer_ | I'm attempting to use LDAP for client authentication. How can I get local users to have the proper groups, and possibly slightly different on a client-to-client basis? I looked at /etc/security/group.conf, but it apparently only supports NIS netgroups, and not LDAP groups. | 20:22 |
| kirkland | jdstrand: where can i get libvirt-*ubuntu26 from? | 20:32 |
| kirkland | jdstrand: i was going to prepare ubuntu27 fixing those two bugs | 20:33 |
| jdstrand | kirkland: unapproved | 20:33 |
| kirkland | jdstrand: got it | 20:33 |
| Dr_Alien | Hi all | 20:34 |
| Dr_Alien | i need some help with installing LAMP | 20:34 |
| Dr_Alien | ive just configured the ftp server however i noticed my apache isnt starting up | 20:35 |
| _ruben | ftp isnt a standard part of lamp usualy ;) | 20:35 |
| Dr_Alien | Yeah :) | 20:36 |
| guntbert | Dr_Alien: what errors do you find in apache's logs? | 20:37 |
| Dr_Alien | Cant see any logs currently bert. | 20:37 |
| Dr_Alien | im a newbie at ssh commands :( | 20:37 |
| _ruben | tail -30 /var/log/apache2/error.log | 20:37 |
| guntbert | Dr_Alien: /var/log/apache2 should contain some | 20:38 |
| Dr_Alien | tail: cannot open `/var/log/apache2/error.log' for reading: No such file or directory | 20:38 |
| Dr_Alien | reimaging my server 1 sec. (on a cloud) | 20:38 |
| Dr_Alien | i used this command. | 20:40 |
| Dr_Alien | sudo tasksel install lamp-server | 20:40 |
| jdstrand | kirkland: cool, thanks :) | 20:40 |
| mean67 | anyone in the boston area | 20:41 |
| _ruben | Dr_Alien: that's supposed to give you a working lamp stack, then again, havent used that "task" for quite some time | 20:42 |
| Dr_Alien | Ok its up now.. | 20:42 |
| Dr_Alien | any easy way of installing a ftp serveR? | 20:42 |
| Dr_Alien | server* | 20:42 |
| Dr_Alien | i want to put html files up to my server quickier than editing them in vi. | 20:43 |
| _ruben | i tend to use pure-ftpd, but others might have different preferences like vsftpd or proftpd | 20:44 |
| _ruben | or even better: scp/sftp | 20:44 |
| mean67 | anyone assist me with a UEC install that I am trying to do | 20:44 |
| Dr_Alien | Ok | 20:45 |
| Dr_Alien | Could you help me install this? | 20:45 |
| Dr_Alien | im a total newbie at installations lol | 20:45 |
| mean67 | well I am just trying to set it up right | 20:46 |
| _ruben | https://help.ubuntu.com/9.10/serverguide/C/ftp-server.html | 20:47 |
| Dr_Alien | hmm interesting | 20:47 |
| Dr_Alien | ok | 20:47 |
| Dr_Alien | 20:47 | |
| Dr_Alien | /usr/sbin/vsftpd already running. | 20:47 |
| Dr_Alien | [fail] | 20:47 |
| Dr_Alien | What would the default FTP username and password be? | 20:47 |
| Dr_Alien | and what could i do to change the username and password | 20:48 |
| kirkland | jdstrand: can you give this little patch an eyeball? http://paste.ubuntu.com/420015/ | 20:48 |
| kirkland | jdstrand: i think it's better (simpler) than the bug reporter's | 20:48 |
| kirkland | jdstrand: and I think we can get it upstream | 20:48 |
| kirkland | jdstrand: a) instead of changing every line that uses $SYSCONFDIR, just set it if unset | 20:48 |
| Dr_Alien | Status:Connection attempt failed with "ECONNREFUSED - Connection refused by server". | 20:48 |
| kirkland | jdstrand: b) change \t to \s* to handle any leading whitespace | 20:48 |
| Dr_Alien | Error:Could not connect to server | 20:48 |
| electro_ | RoyK: Do you have a link that explains the Redundancy issue? | 20:49 |
| kirkland | jdstrand: c) fix the bashism, rather than changing interpretters | 20:49 |
| kirkland | jdstrand: opinion? | 20:49 |
| _ruben | Dr_Alien: by default it most likely uses the same credentials as your local/ssh login | 20:50 |
| Dr_Alien | Ok Thanks | 20:50 |
| Dr_Alien | and what about the port number? | 20:50 |
| _ruben | i'd guess 21, but you can check with netstat -lnt | 20:51 |
| Dr_Alien | Ok | 20:52 |
| Dr_Alien | its saying only annoymas connections can be done | 20:52 |
| Dr_Alien | Response:530 This FTP server is anonymous only | 20:52 |
| _ruben | read the "User Authenticated FTP Configuration" section | 20:54 |
| kirkland | jdstrand: updated, http://paste.ubuntu.com/420019/ | 20:55 |
| kirkland | jdstrand: part of this is fixed upstream already, part of it isn't | 20:55 |
| wack479 | would/should a mounted raid array cause the drive that it is mounted on to show as being "full"? | 20:55 |
| Dr_Alien | where Ruben? | 20:55 |
| Dr_Alien | _ruben where would i find that? | 20:56 |
| SirStan | Any recommendations on a server 'panel' like cPanel? | 20:57 |
| _ruben | Dr_Alien: the url i pasted earlier | 20:57 |
| jdstrand | kirkland: did you test it? does @SYSCONFDIR@ get properly replaced? | 20:57 |
| jdstrand | kirkland: otherwise that last paste seems fine | 20:58 |
| kirkland | jdstrand: going to test, but that's what's currently upstream | 20:58 |
| kirkland | jdstrand: otherwise, see my first paste; that's guaranteed to work :-) | 20:58 |
| jdstrand | kirkland: yeah-- I just don't know if the autoconf bits are in 0.7.5 for that file | 20:59 |
| kirkland | jdstrand: i'll test | 20:59 |
| Dr_Alien | I saw :) | 20:59 |
| jdstrand | kirkland: if it works as desired, go with it, otherwise your first attempt seems fine for lucid | 20:59 |
| Dr_Alien | Ok, how do you save files in nano? | 20:59 |
| kirkland | jdstrand: i've rebased the patch against git and will submit upstream too | 21:00 |
| jdstrand | kirkland: excellent | 21:00 |
| _ruben | Dr_Alien: ctrl+o .. it says so at the bottom of the screen | 21:00 |
| _ruben | anyway im off to bed | 21:00 |
| Dr_Alien | tc and thanks! | 21:00 |
| kirkland | jdstrand: do you care what i number this patch? | 21:01 |
| kirkland | jdstrand: 9023 is "next" | 21:01 |
| kirkland | jdstrand: is that okay? | 21:01 |
| jdstrand | kirkland: 9023 is just fine-- the 9xxx are Ubuntu only | 21:02 |
| wack47 | sry my irc froze, so im gonna ask again... would/should a raid array show the drive it is mounted to as being full? | 21:02 |
| jdstrand | kirkland: before uploading, hold on a sec | 21:02 |
| Dr_Alien | Ok, what would people say is the default password for FTP server? | 21:03 |
| wack47 | password bahahahaha | 21:03 |
| kirkland | jdstrand: sure thing | 21:03 |
| kirkland | jdstrand: debdiff is currently http://paste.ubuntu.com/420024/ | 21:03 |
| kirkland | jdstrand: i'm about to pbuild now | 21:03 |
| wack47 | anyone know about the raid? | 21:09 |
| jdstrand | kirkland: actually, there is a another thing you should change | 21:09 |
| jdstrand | kirkland: echo make sure the gnutls-utils package is installed | 21:09 |
| jdstrand | kirkland: that should be gnutls-bin on Ubuntu | 21:09 |
| jdstrand | kirkland: (and Debian) | 21:09 |
| kirkland | jdstrand: ga-noodles! | 21:10 |
| jdstrand | kirkland: obviously not at all critical, but if inclined, since you are already in there, it would be nice | 21:10 |
| jdstrand | hehe | 21:10 |
| kirkland | jdstrand: | 21:12 |
| kirkland | echo "Could not locate the certtool program" | 21:12 |
| kirkland | echo "make sure the gnutls-utils (or gnutls-bin) package is installed" | 21:12 |
| jdstrand | sure | 21:12 |
| kirkland | jdstrand: hoping that wording might allow this to go upstream too | 21:12 |
| * jdstrand nods | 21:13 | |
| kirkland | jdstrand: cool | 21:13 |
| jdstrand | kirkland: that sed line is not quite right for our version of certtoll | 21:16 |
| jdstrand | certtool | 21:16 |
| kirkland | jdstrand: okay? | 21:16 |
| wack47 | should/would a raid array that is partially filled show the disk (filesystem) as full? | 21:16 |
| jdstrand | kirkland: eg-- the sed looks for: | 21:16 |
| jdstrand | Issuer: CN=Red Hat Emerging Technologies | 21:16 |
| jdstrand | but our certool does something like: | 21:17 |
| jdstrand | Issuer: C=US,ST=TX,...,CN=example.com,... | 21:17 |
| kirkland | jdstrand: okay, so what should the sed look like? | 21:19 |
| jdstrand | working on it | 21:19 |
| kirkland | jdstrand: okay, i'm testing the autoconf change | 21:20 |
| jdstrand | kirkland: sed 's+\s*Issuer: .*CN=\(.*\),.*+\1+' | 21:20 |
| kirkland | jdstrand: cool, thanks, i'll jam that in | 21:20 |
| jdstrand | kirkland: that should work for both formats | 21:20 |
| jdstrand | kirkland: I can make that slightly better I think | 21:21 |
| jdstrand | (I don't like the ',' business | 21:21 |
| jdstrand | ) | 21:21 |
| kirkland | jdstrand: and still maintain RH compatibility? | 21:21 |
| kirkland | jdstrand: essential for getting this upstream | 21:21 |
| jdstrand | kirkland: of course | 21:21 |
| kirkland | jdstrand: /me likes the way you said "of course" | 21:22 |
| jdstrand | heh | 21:22 |
| ziesemer_ | Trying to patch and recompile a PAM module, pam_group. Downloaded the sources, ./configure && make, but I don't have any *.so files - just *.la and *.lo . What am I missing? | 21:25 |
| genii | make install ? | 21:31 |
| kirkland | jdstrand: fyi, i had to fix the SYSCONFDIR a little better | 21:34 |
| kirkland | jdstrand: it can't work at all, in the way that upstream ships it | 21:35 |
| kirkland | jdstrand: ie, $(SYSCONFDIR) can't evaluate | 21:35 |
| wack47 | should/would a raid array that is partially filled show the disk (filesystem) as full? | 21:37 |
| jdstrand | kirkland: *sigh* | 21:37 |
| kirkland | jdstrand: anyway, am i still awaiting your magic sed? | 21:38 |
| jdstrand | kirkland: I tried for a while to use a single sed, but had to punt and do two: | 21:38 |
| jdstrand | sed 's+\s*Issuer: .*CN=++' | sed 's+,EMAIL=.*++' | 21:38 |
| jdstrand | kirkland: feel free to play with it, but it is eluding me | 21:38 |
| kirkland | jdstrand: can't that be: sed -e 's+\s*Issuer: .*CN=++' -e 's+,EMAIL=.*++' | 21:38 |
| jdstrand | kirkland: I need to get back to iso testing :) | 21:38 |
| soren | jdstrand: You know you can just ... | 21:38 |
| soren | kirkland: Yeah what kirkland just said :) | 21:38 |
| * kirkland high fives soren | 21:39 | |
| jdstrand | is -e portable? | 21:39 |
| soren | I believe it is. | 21:39 |
| jdstrand | feel free to use -e and float it upstream then | 21:39 |
| kirkland | jdstrand: hmm, i use it a lot now in byobu, which people have running on dozens of distros and even OSX | 21:39 |
| soren | At least the docs don't say it's a GNU thing, which I find they're usually pretty good at when that's the case. | 21:39 |
| kirkland | jdstrand: ie, i get complaints about shell portability, but that hasn't been one i've heard | 21:40 |
| jdstrand | k | 21:40 |
| jdstrand | sed -e 's+\s*Issuer: .*CN=++' -e 's+,EMAIL=.*++' | 21:40 |
| jdstrand | ^ confirmed to work with both redhat and us | 21:40 |
| uvirtbot | jdstrand: Error: "confirmed" is not a valid command. | 21:40 |
| jdstrand | kirkland: ^ confirmed to work with both redhat and us | 21:41 |
| kirkland | jdstrand: rock | 21:41 |
| kirkland | jdstrand: and i confirmed that autoconf did it's thing | 21:41 |
| jdstrand | kirkland: you might want to mention that this works with gnutls 2.8.5 specifically | 21:42 |
| jdstrand | kirkland, soren: incidentally, what I really wanted to do is a single regex. that is what was eluding me :\ | 21:44 |
| === libertiy is now known as libertiy|away | ||
| kirkland | jdstrand: here's what i'm sending upstream http://pastebin.ubuntu.com/420045/ | 21:54 |
| kirkland | jdstrand: okay, i'm firing off an upload, and the upstream patch | 21:57 |
| jdstrand | kirkland: sounds good, thanks | 21:59 |
| jdstrand | though that regex was a group effort :) | 21:59 |
| kirkland | jdstrand: ;-) | 22:01 |
| wack47 | should/would a raid array that is partially filled show the disk (filesystem) as full? | 22:02 |
| wack47 | hmmm | 22:15 |
| omar8102 | hola a todos | 22:20 |
| wack47 | hello | 22:21 |
| === libertiy|away is now known as libertiy | ||
| wack47 | should/would a raid array that is partially filled show the disk (filesystem) as full? | 22:28 |
| wack47 | should/would a raid array that is partially filled show the disk (filesystem) as full? | 22:28 |
| wack47 | oops | 22:29 |
| ziesemer_ | I wouldn't think so. What are you using to view? df ? | 22:31 |
| === libertiy is now known as libertiy|away | ||
| osmosis | When I lose contact with my ubuntu server at the colo, I have to call the colo and ask for a reset. Is there any way for me to diagnos the cause of the system outage? Appears to be a crash, but without console access before the reboot...i have no idea what the cause is. | 23:14 |
| flyback | osmosis, what country you in? | 23:15 |
| osmosis | flyback, usa | 23:15 |
| flyback | ah is it your own physical server? | 23:15 |
| osmosis | flyback, yes | 23:15 |
| flyback | you might want to consider one of these if you can spare a pci slot and can hold a full length | 23:16 |
| osmosis | flyback, are you gonna say IPMI ? | 23:16 |
| flyback | http://webdevsys.com/lightsOut.htm | 23:16 |
| flyback | yeah | 23:16 |
| flyback | well not ipmi necessarily | 23:16 |
| flyback | is it a brand name box | 23:17 |
| osmosis | flyback, i actually have an IPMI card..but i cant configure it because its not supported by ubuntu hardy kernel. Once i upgrade to lucid, it will likely be supported. | 23:17 |
| flyback | actually | 23:17 |
| flyback | you should be able to use the card anyways | 23:17 |
| flyback | it should have it's own cpu, ram etc | 23:17 |
| flyback | you just won't be able to monitor hw resources or console | 23:18 |
| flyback | but you would be able to power cycle or reboot | 23:18 |
| flyback | also if it has a virtual serial port on it you could tell the console to use that | 23:18 |
| flyback | console=ttyS1,38400 | 23:18 |
| flyback | etc | 23:18 |
| flyback | also if you configure it to support acpi power button event and you configure your linux to reboot if some tapes the power button | 23:19 |
| flyback | should all be possibly without needing a direct driver | 23:19 |
| flyback | anyways.... just some ideas | 23:20 |
| flyback | AAAAAAAAA | 23:22 |
| flyback | brb | 23:22 |
| * flyback goes to numb his dying tooth | 23:22 | |
| mathiaz | smoser: hi! | 23:33 |
| smoser | hi | 23:33 |
| mathiaz | smoser: I'm trying to detach a device using virsh | 23:33 |
| mathiaz | smoser: and run into the following issue: http://paste.ubuntu.com/420077/ | 23:33 |
| smoser | mathiaz, i dont know. i haven't tried using virsh to detach. | 23:34 |
| smoser | i've tested in kvm in the console | 23:34 |
| mathiaz | smoser: how do you detach usually? | 23:34 |
| smoser | and with eucalyptus | 23:34 |
| smoser | i rarely use libvirt | 23:35 |
| mathiaz | jdstrand: I'm trying to detach a device using virsh | 23:35 |
| mathiaz | jdstrand: and run into the following issue: http://paste.ubuntu.com/420077/ | 23:35 |
| smoser | mathiaz, so, eucalyptusdoes do this | 23:35 |
| smoser | other than its not virtio | 23:36 |
| smoser | its scsi | 23:36 |
| smoser | so that could actually be the difference | 23:36 |
| mathiaz | smoser: right - that being said I'm using an older version of libvirt | 23:36 |
| mathiaz | smoser: so the lucid version of libvirt may be fixed | 23:36 |
| smoser | oh, then easy answer | 23:36 |
| smoser | upgrade | 23:36 |
| smoser | :) | 23:36 |
| mathiaz | smoser: :) | 23:36 |
| smoser | mathiaz, but basically it woudl appear to be a bug in the libvirt screen scraping / expectation of kvm | 23:38 |
| mathiaz | smoser: libvirt screen scraping? | 23:39 |
| smoser | libvirt screenscrapes kvm console | 23:39 |
| smoser | s/console/monitor/ | 23:39 |
| smoser | err... whatever its called | 23:39 |
| smoser | the thing that you would type into if you didn't use libvirt | 23:39 |
| persia | Does it really reinterpret SDL output, or does it tell KVM to provide a text stream? | 23:41 |
| smoser | well it doesn't do graphics to text | 23:42 |
| smoser | it gets a text stream | 23:42 |
| smoser | but it still 'scrapes' | 23:42 |
| smoser | ie, the output changes, its not an api to kvm | 23:42 |
| jrwren | in the topic, doc and resources link says http://tinyurl.com/ubuntuserv but tinyurl says it can't find sites url to redirect to | 23:46 |
| genii | jrwren: I believe it may have pointed to: https://help.ubuntu.com/9.10/serverguide/C/index.html previously | 23:49 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!