[00:02] hggdh: status check [00:02] New bug: #562635 in krb5 (main) "Sync krb5 1.8.1+dfsg-2 (main) from Debian unstable (main)" [Wishlist,Confirmed] https://launchpad.net/bugs/562635 [00:07] kirkland: finalising installs [00:10] kirkland: only waiting for the two NCs to finish install. UEC image is registered [00:11] hggdh: k [00:11] mathiaz: nearly done .... [00:21] kirkland: I need to jet out for now [00:21] mathiaz: are you traveling all day tomorrow? [00:22] kirkland: yeah - I'll be able to retrieve/read/respond to email though [00:22] kirkland: (when I'm not on planes) [00:22] mathiaz: okay, we'll try to get the debug info to you [00:22] kirkland: the next step is to run the tests again and save the log information [00:23] kirkland: put them online somewhere if they fail again [00:23] mathiaz: what specifically do you want? a tee of the run ? [00:23] kirkland: the whole test run log [00:23] hggdh: ^ [00:25] kirkland: nodes registered, running a single-instance test now [00:33] kirkland: test running, log is being written to ~/uec-testing-scripts/resutls/single* [00:33] kirkland: on cempedak [00:33] hggdh: cool, and you can ssh in? [00:35] kirkland: negative [00:35] hggdh: cannot ssh in [00:35] kirkland: ssh fails on timeout [00:35] really sounds like routing [00:36] hggdh: interesting [00:36] hggdh: okay, put the log somewhere for me to check out [00:38] kirkland: k. I just ran one instance by hand, and then tried to ssh into it -- fails with a timeout [00:39] hggdh: okay, that's easy to reproduce [00:39] hggdh: log? [00:42] kirkland: people.c.c/~cerdea/single_test.log.2010-04-13_193218 [00:46] hggdh: rsync -aP people.canonical.com:~cerdea/single_test.log.2010-04-13_193218 . [00:46] hggdh: file not found [00:47] hggdh: found it, public_html [00:47] heh. one wants it on public_html, another on the root ;-) [00:49] hggdh: ls -alF users/admin/uectest-k0.priv [00:50] hggdh: and cat that file, make sure it matches -----BEGIN RSA PRIVATE KEY----- [00:50] hggdh: is that instance still running? [00:50] hggdh: can you telnet to its port 22 ? [00:51] kirkland: yes, the instance is still running [00:52] kirkland: the priv key seems kosher [00:52] hggdh: and telnet ? [00:53] kirkland: timeout. Also, a traceroute (FWIW) reaches marula (the CC) and stops there [00:54] hggdh: oh, interesting [00:54] hggdh: that's got to be it [00:54] kirkland: let me try to ssh from marula [00:54] hggdh: yeah [00:54] hggdh: scp the priv key over [00:54] hggdh: and try from there [00:55] kirkland: first test -- reachability -- successful [00:55] hggdh: ack [00:55] will move the priv key there now [01:00] hggdh: and? [01:00] kirkland: getting permission denied (pub key) [01:00] kirkland: but the important piece is that I am *reaching* the instance [01:00] hggdh: hrm, odd [01:00] hggdh: agreed on that point [01:00] hggdh: and you're doing ssh -i ./whatever.priv ubuntu@ip ? [01:00] hggdh: and whatever.priv is perm'd 600 [01:01] kirkland: yes indeed, and will check again [01:01] but on wrong permission ssh would bail out [01:03] kirkland: and the full command is ssh -o UserKnownHostsFile=/dev/null -o StrictHostKeyChecking=no -i ./uectest-k0.priv ubuntu@10.55.55.100 [01:04] although sort of overworked, I admit [01:04] hggdh: hmm, okay [01:04] hggdh: it may be that the guest is having trouble getting out [01:04] hggdh: or at least to have the key injected [01:04] hggdh: okay, add your traceroute findings to that bug [01:05] hggdh: and email mathias (cc me) the link to that log [01:05] hggdh: i'm reassured that this appears to be a networking issue, but we'll need to get to the bottom of it [01:05] hggdh: i gotta run for the night [01:05] hggdh: thanks dude! [01:05] kirkland: will do, and g'night [01:06] set up key based auth with ssh, but still get password prompts on one of my computers... any idea why? [01:38] What's the best/easiest way to share a directory on a secondary drive with no password access for those on my LAN? === RoAk is now known as RoAkSoAx [03:48] who is the server team manager? [03:48] as in, job description, Reports To: Server Team Manager [03:49] Maybe https://launchpad.net/~ubuntu-server says [03:49] It's "owned" by mathiaz, for whatever that's worth. [03:50] (I tend to avoid lp, so I'm just guessing.) [03:50] He's not the server team manager === smoser_ is now known as smoser [04:18] Can anyone here help me with a few basic questions concerning OpenLDAP under 9.10 or 10.04? [04:19] !anyone [04:19] A large amount of the first questions asked in this channel start with "Does anyone/anybody..." Why not ask your next question (the real one) and find out? [04:19] OK, I was asking for that. [04:20] So dpkg-reconfigure no longer really does anything for OpenLDAP in Karmic in newer. [04:20] I.E., LDAP is pretty much unusable once installed, apparently without a large amount of additional configuration. [04:20] Two such guides I found for this are at http://ubuntuforums.org/showthread.php?t=1313472 and http://ubuntuforums.org/showthread.php?p=8154148 . However, they vary quite a bit, etc. === dendrobates is now known as dendro-afk [04:21] Is there anything more official towards getting this running other than a forum post? [04:22] For example, the first one uses back_hdb for olcModuleLoad, the other uses back_bdb.la . What's the difference? [04:22] ziesemer_: have you tried the ubuntu server guide? [04:23] https://help.ubuntu.com/9.10/serverguide/C/openldap-server.html ? Yes, but apparently it is very wrong and incomplete, as detailed in those forum posts. [04:24] Are you running 9.10 or 10.04? [04:24] Either. I played with this previously on 9.10. I'm trying to start from scratch on a 10.04 VM, since it'll be official in just a few weeks... [04:25] Have you looked at the 10.04 ubuntu-serverguide? [04:25] I couldn't find. Google wasn't helping... [04:26] Was at least suspecting it wasn't done yet? [04:26] apt-get install ubutu-serverguide in whatever you're running [04:27] Oh - not available online? [04:27] I don't know if it's going to help, I'm just giving you the standard triage [04:27] np [04:29] Installed - how do I access? [04:29] No man page, etc. [04:30] ziesemer_: dpkg -L ubuntu-serverguide [04:30] should tell you what files were installed where [04:31] it's probably some HTML files in /usr/share/doc/ubuntu-serverguide/ [04:31] file:///usr/share/ubuntu-serverguide/html/C/index.html - thanks. [04:31] New bug: #562746 in php5 (main) "Apache2 wont start, missing lib - 10.04" [Undecided,New] https://launchpad.net/bugs/562746 [04:32] Interesting. They pretty much just copied one of the forum posts into the doc. [04:36] New bug: #562750 in apache2 (main) "Apache2 wont start, missing lib - 10.04" [Undecided,New] https://launchpad.net/bugs/562750 [04:43] anyone know of a csv to html tool that takes multiple columns and generate a html table and a bar chart ? [04:50] gnuplot? === jldugger is now known as pwnguin [04:53] twb: gnuplot .. ok ... let me check it out === dendro-afk is now known as dendrobates [05:02] I don't know if gnuplot can emit the format you want specifically. [05:02] But it's the obvious way to turn data into graphs/charts. [05:07] twb: i am new to it.. have not fig out how to plot yet ;-) [05:07] It's not exactly intuitive :-P [08:30] Hi, I have just installed a new machine with user's home encrypted. Now I need to run usermod -u but this not affect the unmounted home partition [08:31] can I mount the user's home from root? [09:06] New bug: #562832 in drbd8 (main) "module drbd8 update kernel from 2.6.32-16 to 2.6.32-20" [Undecided,New] https://launchpad.net/bugs/562832 [10:07] I'm trying to get an email-server working on my Ubuntu 9.04 VPS. Checking my mail.log file, I see Apr 14 09:59:13 vps2735 postfix/master[28009]: fatal: bind 127.0.0.1 port 10024: Address already in use - Port 10024 is being used by Amavisd... what should I do? [10:09] To be honest, I really don't care about Amavisd - all I want is a simple email server up and running. The tutorials I followed(or tried to) had other ideas, however... [10:29] <_ruben> RobbieThe1st: then dont use amavisd [10:30] <_ruben> regarding the use of tutorials, the #postfix channel bot has a clear opinion abou tit [10:30] <_ruben> :30 < knoba> _ruben: "tutorial" : A very common problem is that some people prefer to follow a step-by-step tutorial that shows them how to setup their server w/out reading the documentation or understanding what they are doing. If [10:30] <_ruben> something goes wrong, they have no clue whtsoever about where to find hints, and they sometimes decide to start from scratch using a different tutorial. This is not Th e Proper Way. [10:31] I'm trying -yet again- to uninstall it, yet for whatever reason, last time I did it Postfix was looking for Amavisd. [10:32] And, no its not the true way. Or the best way, but I can't understand all of the documentation, so I figure the best thing to do is to use tutorials until I get things mostly working, then experiment from there on my own [10:35] I'm wondering if the http://ppa.launchpad.net/ubuntu-ha/ is going to be fully implemented in the standard Ubuntu repo, or do I need to enable this to be running get the complete cluster stack? [10:36] *-get* [10:51] Good day. In #ubuntu-powerpc we've been discussing oversized CDs. The Ubuntu sever CD is one of those that is oversized on powerpc. I wondered if the eucalyptus stuff could be safely dropped from powerpc because kvm doesn't support powerpc and we don't ship the IBM hypervisor. [10:52] If that's not enough (it ought be), I wonder if anyone would object to dropping other virtualisation bits (and would hope someone could give me a list of good candidates) [11:06] New bug: #562912 in munin (main) "munin-node plugin initialisation fails on ip_" [Undecided,New] https://launchpad.net/bugs/562912 [11:26] New bug: #562919 in samba (main) "Lucid Beta2: Desktop Crash, then Unable to Boot" [Undecided,New] https://launchpad.net/bugs/562919 === jussi01 is now known as jussi [12:41] persia: yes, that sounds like a good candidate for removal [12:42] persia: however the eucalyptus.udeb might be linked from the "Install UEC" option on that CD [12:42] persia: you probably would need to get rid of that as well [12:43] ttx: How is the "Install UEC" option defined? [12:43] persia: I think it's directly on the ISO itself, let me check [12:44] My worry is that this may not be an arch-dependent-aware construction. [12:44] persia: right, that's why I mention it [12:44] But I am fairly sure that attempting to do a powerpc cloud without IBM's hypervisor is doomed to failure :) [12:45] syslinux/text.cfg has "menu label Install Ubuntu ^Enterprise Cloud" [12:45] (unless someone steps up and ports another hypervisor, but given that only a small proportion of hardware supports it, and all of that comes with IBM service contracts, ...) [12:45] on the ISO itself [12:45] then it points to /cdrom/preseed/cloud.seed [12:46] * persia pulls debian-cd to see how that is constructed [12:46] ttx: Would you mind adding [!powerpc] to the relevant packages in the seed anyway? I'll try to sort out how to make the option go away, but I'd rather a CD with a bug than an oversized CD. [12:48] persia: that would be for package eucalyptus only ? (and then the deps will not get pulled ?) [12:49] * persia looks at the seed [12:49] persia: kirkland should upload a new eucalyptus today, could you sync with him when he is up ? [12:50] new eucalyptus? [12:50] ;) [12:50] well, cherrypicking the last bugfixes from upstream branch [12:52] ok, just wondering because apt-get installed already new ones yesterday [12:52] didn't think it updates that often..... [12:52] Ugh. Looks like it might be complicated, because the eucalyptus stuff is all split out in other seeds, which would need to be completely [!powerpc]'d [12:52] Or maybe better to have them [i386 amd64] ? Does eucalyptus work on sparc/ia64? I know it doesn't on powerpc/armel (no kvm). [12:53] ttx: But, yeah, I'll try to catch kirkland when it's daytime there. [12:53] he should be around in ~1h [12:53] I don't think it works on sparc/ia64 [12:54] Then, yeah, we ought special-arch that stuff. [12:54] ttx: Do you know how to pass kvm parameters when running instances? I want the kvm instances launched with -vnc [12:54] it builds, but I suspect kvm will play weird games [12:55] binBASH: no, maybe ask in #eucalyptus [12:55] Well, we build qemu-kvm, so we end up with working qemu even when kvm isn't around. [12:55] ttx: Trying since some days now, I will ask in their forum [12:55] But yeah, I don't know enough about hypervisors for sparc/ia64 to know if it *can* work. [12:56] persia: I think it depends on the cpu features. [12:56] And ia64 CD is also oversized (although I don't personally care), so making it [i386 amd64] may help there as well. [12:56] not on the hypervisor itself [12:56] binBASH: Well, yes, but also on the implementation. For example, kvm hasn't been ported to use PowerVM [12:57] (mostly because most folks who get hardware that has this feature are happy to get IBM's lower-level utilities and use that) [13:22] Does anyone here know if keepalived in Lucid supports ipv6? [13:22] Can't find anything about it [13:31] mvo: ping [13:33] ttx, ping [13:33] smoser: yo [13:33] so the libsasl2-2 issue i mentioned. [13:33] thats not really a bug, right, but an unsupported upgrade ? [13:33] New bug: #562960 in samba (main) "folder sharing in gnome not working in Lucid" [Undecided,Incomplete] https://launchpad.net/bugs/562960 [13:34] smoser: I suspect it doesn't happen on a true do-release-upgrade [13:34] so it's taken care of by the upgrade script [13:34] you are correct, it does not happen there. [13:34] thats what i thought. [13:35] smoser: then no, it's not a bug, it's an unsupported scenario that needs some effort to work [13:36] smoser: we'll let you test a little your latest, give a chance to mathiaz to ack, and I can sponsor it EOD in any case [13:36] well i just tested the hardy2lucid [13:36] which is the only case in which the code i most recently changed fires [13:36] hardy2lucidppa direct ? [13:36] yeah [13:37] yes, I tested that one a few mion. ago [13:37] got my email ? [13:37] right. i saw your mail and responded. [13:37] ok [13:37] as much as i can tell its right. [13:37] it was a lot simpler once i understood what was needed.... [13:37] To test "real upgrade" we'ell have to get it into the archive anyway [13:38] cn=config is on the long list of things I've got to find time to play with, unfortunately nowhere near the top [13:38] your explanation in the bug was good, but having no familiarity with it, i was misled a bit. [14:36] zul: mvo is not on this channel [14:50] help.ubuntu.com down? [14:51] binBASH: its certainly taking a long time to access here. [14:52] Pici: I get here 503 [14:52] Service Temporarily Unavailable [14:52] The server is temporarily unable to service your request due to maintenance downtime or capacity problems. Please try again later. [14:52] Apache/2.2.8 (Ubuntu) mod_ssl/2.2.8 OpenSSL/0.9.8g Server at help.ubuntu.com Port 443 [14:55] binBASH: I'm poking some people about it. [14:55] persia: howdy [14:56] New bug: #563039 in bacula (main) "package bacula-director-mysql 5.0.1-1ubuntu1 failed to install/upgrade: subprocess installed post-installation script returned error exit status 1" [Undecided,New] https://launchpad.net/bugs/563039 [14:57] Pici: Now I get some Squid error ;) [14:57] kirkland: Good day. So, the powerpc CD is oversized, and I'd like to cut some stuff. Am I correct that eucalyptus wants kvm, or does it also work against qemu? [14:57] binBASH: Try again, it just started working for me. [14:58] It's back! thx Pici === dendrobates is now known as dendro-afk [15:08] hi, if I use lftp ftp://login:pwd@ip -e "cd / ; mirror -R /DIR_DIST_BACKUP ; quit" for save any ubuntu server, is there any risk? [15:21] New bug: #563053 in mysql-dfsg-5.0 (universe) "Please remove Mysql 5.0 from the archive for lucid." [Undecided,New] https://launchpad.net/bugs/563053 === dendro-afk is now known as dendrobates [15:37] neuro666: risk in what way? In the clear password presents some risk perhaps. [15:38] Hi all. I'm trying to configure my second NIC, I'm doing it via editing the file /etc/network/interfaces. Jere [15:38] *crap* [15:38] anyone using CUPS with SAMBA to store Windows drivers? [15:41] I'm trying to configure my second NIC with a public IP. I'm editing the file /etc/network/interfaces. I've got my public IP, netmask, and gateway that I need to assign. But do I need to specify a 'network' and 'broadcast' line in the config? What are those for? How do I find out what addresses to assign those if they are required? [15:44] <_ruben> b14ck: not required, they are automagically determined based on ip+netmask [15:44] b14ck: What you want to do is not difficult. You just type your netmask, gateway, network, broadcast and address under the 'iface eth? inet static' line. See http://www.debian-administration.org/article/An_introduction_to_Debian_networking_setup [15:44] Eh, as _ruben said. They are not required [15:45] Ah thanks you both. I'll read that guide too. [15:45] Appreciate it! [15:56] New bug: #563084 in bacula (main) "bacula-director-pgsql does not install cleanly with db-config" [Undecided,New] https://launchpad.net/bugs/563084 [15:57] kirkland: So, ttx suggested that if eucalyptus does depend on kvm, it would be safe to drop it from the CD. The menu adjustments in debian-cd seem already to be i386/amd64 specific. What do you think about adding [i386 amd64] liberally to the euca* seeds? [15:58] Now if eucalyptus truly depends on kvm, it may make sense to set Architecture and P-a-s it, but if it works with qemu, maybe we can do a seed-only change. [15:59] ttx: persia: the only part of eucalyptus that actually depends on kvm is eucalyptus-nc (the node controller itself) [15:59] persia: that said, i don't have any interest in support eucalyptus on any architecture other than amd64 [15:59] zul: hi! [16:00] jdstrand: hey! [16:00] kirkland: Not even i386? [16:00] persia: personally? no. but we shouldn't drop i386 :-) [16:00] zul: so on bug #559628 I think it should be marked "Won't Fix" per my comment #3 [16:00] Launchpad bug 559628 in ntp "ntpd profile denies access to /etc/ld.so.preload" [Low,Triaged] https://launchpad.net/bugs/559628 [16:00] persia: i think it would be acceptable to force UEC to amd64|i386 in the CD seeds [16:01] persia: and leave the builds in the archive for ambitious arm,powerpc,sparc users [16:01] zul: basically, I think the apparmor profile is doing its job [16:01] jdstrand: sounds good to me [16:01] kirkland: OK. I don't have write access to the server seed. Would you mind adding "[i386 amd64]" to the relevant packages in those seeds? [16:02] persia: (the reason for my amd64 comment ... amd64 hardware can host both 32bit and 64 guests; i386 hardware can only host 32bit guests [16:02] kirkland: Also, eucalyptus-nc probably wants to Recommend: qemu-kvm rather than "kvm" to avoid installing the transition packages on new installs. [16:02] Makes sense. [16:02] persia: good call [16:02] Thank cjwatson for it, really. [16:02] persia: i'll do it, if you'll double check my work for me before commit [16:02] ttx: are you okay with this? [16:03] zul: do you mind 2nding my opinion and updating the bug then? [16:03] kirkland: Be happy to do so, thanks. I believe this will get us burnable powerpc and ia64 CDs again (although we may need to find something else to trim). [16:04] persia: eucalyptus-nc recommending qemu-kvm committed [16:04] jdstrand: sure [16:04] zul: thanks [16:05] jdstrand: done [16:07] What package is responsible for changing entries in /etc/fstab to UUID on upgrade? I'd like to file a bug against it. [16:11] kirkland: yes [16:11] cjwatson: persia: will this suffice? http://paste.ubuntu.com/414373/ [16:12] kirkland: that's easily reversible if it breaks the world as we know it [16:13] I'd recommend [amd64 i386]'ing the entire tasks, not just eucalyptus-* [16:13] otherwise you have weird half-empty tasks on other architectures [16:13] cjwatson: Can that be done in STRUCTURE? [16:13] RoAkSoAx: around ? [16:13] Hi guys, one of my network interfaces isn't working, but I can't figure out why [16:14] What command can I use to verify that my physical NIC port is UP and should start using the statically assigned IP when the ethernet port is plugged in? [16:14] persia: no, sorry [16:14] cjwatson: hmm, i thought about that, but this didn't show me how: grep "Task.*i386" * [16:14] kirkland: I meant just every package in those tasks [16:14] cjwatson: ah [16:14] cjwatson: I hadn't thought so, but figured it was worth asking :) [16:16] So users will end up with empty tasks for ports, but that's fine, as if they want to install UEC, they ought be doing it on i386/amd64 anyway. And if someone wants to fix it by getting it to work nicely on another arch, the packages are available, and we can change in the future (assuming there's space on the CDs) [16:17] persia: cjwatson: http://paste.ubuntu.com/414376/ [16:17] kirkland: Did you mean to drop screen from all the lsits? [16:17] (note that i also dropped screen, which is no longer necessary since we have eucalyptus-udeb now installing the server seed too) [16:17] persia: yup [16:18] In that case, yeah, looks fine. [16:18] persia: i can do this in 2 separate commits, if you like [16:18] Doesn't matter to me, I just wanted to verify it was intentional :) [16:18] persia: yeah, thanks [16:19] persia: cjwatson: r1699 pushed [16:20] I would really know the name of the package that is responsible for the change to UUID in /etc/fstab. Is it update-manager-core? [16:20] cjwatson: ttx: can we get a server cd build going for i386 and amd64 just to make sure this is good? [16:20] Excellent. Does that need a -meta upload also, or should I see the effects on the next round of ports CDs? [16:20] alvin: totally guess here, but i'd expect mountall [16:21] If we're building ports CDs, can we do powerpc and ia64 too? I want to make sure they get un-OVERSIZED from this. [16:21] persia: i don't know the answer to that question [16:21] kirkland: not right now, needs a d-i upload first [16:21] persia: it needs two non-empty publisher runs [16:22] so I recommend against rushing to do CD spins in general for this, it will probably only confuse [16:22] kirkland: Thanks [16:23] Right then. I'll check at normal rebuild time. [16:26] is there any other software to test hd performance other than iozone or bonnie ? [16:29] ttx, i am [16:31] RoAkSoAx: about cluster stack status [16:31] ttx, well everything is in universe besides corosync [16:31] i mean cluster-agents, cluster-glue, pacemaker and heartbeat [16:31] RoAkSoAx: did you file the sync requests [16:31] ? [16:32] with debian ? [16:32] I'm trying to automate install of ubuntu server on many machines using kickstart. All steps go fine except that setup complains about missing default route and needs user input.how do I bypass this? [16:32] ttx, i did [16:32] RoAkSoAx: and they were taken care of yet ? [16:32] ttx, im waiting for FFe to be accept to be able to upload the latest packages [16:32] RoAkSoAx: ok, maybe ping ScottK, would be good to have before the Freeze [16:33] ttx, i already did :) [16:33] RoAkSoAx: ok [16:33] RoAkSoAx: anything we can do to help ? [16:33] ttx, other than that, ivoks and I discussed what to do, and we decided to keep some packages in ubuntu-ha-maintainers and provide updates there.. and then I guess we could backport them [16:34] ttx, other than that.. I think we are pretty much all set with the new clsuter stack [16:35] RoAkSoAx: sounds good to me [16:36] ttx, and I guess we'll try to SRU docs for the server guide [16:36] once polished [16:37] RoAkSoAx: good wikidocs are ok anyway [16:37] * ttx brb [16:40] ttx, ok then. We still have some work items left that I guess could be worked for maverick [16:41] RoAkSoAx: ok, feel free to mark them "POSTPONED" on the blueprint whiteboard [16:41] New bug: #563114 in lm-sensors (universe) "w83697hf-isa-0290 wrong voltage values and tags" [Undecided,New] https://launchpad.net/bugs/563114 [16:41] ttx, will do [16:46] zul, kirkland woudl one of you be able to sponsor a openldap upload later today? [16:46] smoser: sure [16:46] smoser: sure [16:46] we're hoping to get mathiaz review and sponsor, but if he doesn't come around then we'll need someone else [16:46] smoser: lemme know when [16:47] we'll give mathiaz till 4:00 PM US/Eastern (20:00 UTC) [16:57] kirkland: Do you know how to pass kvm parameters when running instances? I want the kvm instances launched with -vnc [17:00] binBASH: through libvirt? [17:00] binBASH: if you're using libvirt, you need to edit the xml, or use virsh [17:01] kirkland: yup, I mean in eucalyptus btw. Didn't find any xml :( [17:02] binBASH: i'm not sure, i haven't tried [17:02] kirkland: http://forum.eucalyptus.com/forum/kvm-inbuilt-vnc [17:02] this is what I get :/ [17:02] What's the best way to force 'ssh' to start at boot? [17:02] Mine doesn't start automatically for some reason. [17:04] b14ck: you'll need to figure out why. does manually running the init script work? [17:04] bladernr: /etc/init.d/ssh starts it, and /etc/rc2.d/Sxxsshd should be a symlink to it [17:05] RoyK: huh? [17:05] binBASH, are you looking to hack ? [17:05] i can hack this for you [17:06] smoser: Just want to get it running somehow ;) [17:06] /etc/init.d/ssh starts it just fine [17:06] well, on the NC there is /usr/share/eucalyptus/gen_kvm_libvirt_xml [17:06] bladernr: heh - s/bladernr/b14ck/ [17:06] that is the thing that writes the libvirt xml [17:06] I was thinking I had to run update-rc.d or something [17:07] b14ck: no, you should not need to do that with ubuntu [17:07] modify it, and put a vnc console stanza in, and then you will (untested) be able to get there via vnc. [17:07] the difficulty will be in knowing which of the NC has the instance you just launched [17:07] it is possible with log scraping on the CC [17:07] pmatulis, when I do an ls -la /etc/init.d | grep ssh, it is not a symlink [17:08] b14ck: no, the symlink is in /etc/rc2.d [17:08] smoser: That was indeed a very good hint, will try it later. [17:08] thanks a lot [17:09] oh, ya that symlink is there [17:09] Ssomethingssh [17:11] smoser: Within the script it's also possible for me to force mac address based on nodes hostname. Very good ;) [17:11] well, ... maybe. you wont have much information to go on, though [17:11] but it might work. i'm not sure. [17:12] I will try ;) [17:19] binBASH, http://pastebin.com/B0A8K3Zc [17:19] i did that, started an instance and verified connection [17:19] however, you can't log in because ubuntu user and root doesn't have a password (in the UEC images) [17:22] you'd have to set one . that can be done scripted with 'echo ubuntu:ubuntu | chpasswd' [17:22] or, possibly more clearly 'echo ubuntu:newPassWord | chpasswd' [17:26] Hi [17:30] ttx: so yes basically Bug #542208 is the same and we need to blacklist the ATI ES1000 from kms [17:30] Launchpad bug 542208 in linux "Please blacklist i830 from Kernel mode-setting" [Critical,In progress] https://launchpad.net/bugs/542208 [17:35] anyone know how to set internal-sftp users umasks? [17:41] v0lksman: create a new login class i guess [17:41] just read that somewhere...how do you do that? [17:41] (docs are fine...just no clue about that) [17:41] v0lksman: man 5 login.conf [17:41] yeah: No manual entry for login.conf in section 5 [17:43] v0lksman: i'm sure you'll find what you need eventually === arm is now known as Guest86001 [17:46] "You" here is the upstream developers of OpenSSH, and they really did [17:46] mean login.conf(5), since they're working on OpenBSD. Indeed it doesn't [17:46] exist on Linux, though. [17:47] i just need to make sure that when an sftp user uploads a new file it is group writable... [17:49] v0lksman: maybe a wild cron job? [17:49] hhaha...thought about it...but that just seem wrong... === swift_ is now known as swift [17:50] v0lksman: yeah [17:50] openssh 5.4 has the ability built in...and there is a patch available for 5.1 but again seems wrong to have to go that route...I think I'm stuck [17:52] v0lksman: run an OpenBSD server? ;) [17:52] hahaha...almost looking that way...I already had to break my LTS to get openssh5.1 so I could easily add chroot'd sftp users now this! :) [17:53] v0lksman: cjwatson is here and he is the maintainer. maybe get his thoughts [17:54] thanks! hopefully he will chime in... :) [18:02] MTecknology: approve my linked in group application! [18:03] v0lksman: i found this: Subsystem sftp /bin/sh -c ‘umask 0002; /usr/libexec/openssh/sftp-server’ [18:04] pmatulis: I don't think that works with 5.1...that's for older versions of openssh (< 4.6) if I'm correct [18:05] I bit the bullet and applied the patch to 5.1...not happy about it but it works... [18:05] going to open a bug to see if it can be added in a backport or something...I hate having to maintain ssh on my own [18:06] New bug: #563202 in euca2ools "euca2ools often output error messages to stdout" [Undecided,New] https://launchpad.net/bugs/563202 [18:06] v0lksman: so you made your own package? [18:06] yep [18:06] v0lksman: ok [18:06] http://sftpfilecontrol.sourceforge.net/ using that patch [18:07] very simple patch...I think it got approved by openssh and is now part of 5.4 [18:21] New bug: #563216 in openssh (main) "Openssh5.1p1 sftp file control" [Undecided,New] https://launchpad.net/bugs/563216 [18:22] smoser: thx for patch, I'll try that [18:22] New bug: #551901 in krb5 (main) "likewise-open fails to join Windows 2000 SP4 domain" [Undecided,Confirmed] https://launchpad.net/bugs/551901 [18:24] Don't you love when your trying to help someone in an irc channel, and they just part without saying anything. === dendrobates is now known as dendro-afk === dendro-afk is now known as dendrobates [18:40] So, I know this has been asked before, but does anyone know what the difference between a command line only install and a ubuntu server install? [18:41] Other than the ubuntu-server kernel of course. === dendrobates is now known as dendro-afk === nxvl_ is now known as nxvl [18:51] pwnguin: please read directions [18:53] pwnguin: If I wasn't clear then please let me know. However, I won't approve it until you follow that one simple step. [18:56] New bug: #563240 in mysql-dfsg-5.1 (main) "package libmysqlclient16 5.1.41-3ubuntu12 failed to install/upgrade: error writing to '': Success" [Undecided,New] https://launchpad.net/bugs/563240 [19:01] New bug: #563241 in mysql-dfsg-5.1 (main) "package mysql-server-5.1 5.1.41-3ubuntu11 failed to install/upgrade: subprocess new pre-removal script returned error exit status 1" [Undecided,New] https://launchpad.net/bugs/563241 [19:11] exit [19:13] smoser: I have to make this changes on cluster controller I think? [19:13] node controller [19:13]