2008/02/09/#ubuntu-server.txt

[03:52] <c1|freaky> does anyone know of a real good groupware? (just not eGroupware?)
[04:01] <soulc> ummmm yeah just dumped fc7 for ubuntu server
[04:02] <soulc> proper install for web server/ mysql php etc is LAMP?  and will I still have the mail options?
[04:08] <faulkes-> soulc: you can choose a lamp install that will include mail
[04:08] <faulkes-> typically that being postfix
[04:09] <soulc> good I am tired of sendmail
[04:09] <faulkes-> awww, come on now, whats not to love about m4 and mc files ;)
[04:11] <soulc> ha ha ha
[04:11] <soulc> tired of the spoofed spam just so I can have email
[04:12] <soulc> so I am reinstalling
[04:14] <soulc> how do I tell the installer that I don't want dhcp
[04:49] <ScottK> IIRC have dhcp not be available.  It's easy enough to reconfigure later, I wouldn't sweat it.
[04:51] <soulc> so how do I configure network settings then I got it up and runnering
[04:53] <ScottK> By editing /etc/network/interfaces
[04:53] <soulc> and where is the cgi-bin dir?
[04:53] <soulc> ok
[04:54] <soulc> thanks I am used to fc
[04:54] <ScottK> No problem.
[04:54] <ScottK> I think /var/www somewhere, but I'm a mail server guy.
[04:56] <soulc> auto eth0 is running dhcp right
[04:56] <ScottK> As far as mail servers go, Postfix is what is primarily supported, but packages for all the popular open source MTAs are in the repositories.
[04:57] <ScottK> Yes
[04:57] <soulc> so if I set it as inet
[04:58] <ScottK> Here's a working static config http://paste.ubuntu-nl.org/55309/
[04:58] <soulc> thanks
[05:10] <lamont> ScottK: still awake?
[05:10] <ScottK> Yeah
[05:11] <lamont> 345047 == please discover ipv6 configured networks
[05:11] <lamont> sigh
[05:11]  * ScottK looks
[05:11] <ScottK> lamont: I thougth we tested this one and it was fine.
[05:12] <lamont> ah, /me looks at the debdiff
[05:12] <ScottK> That's the one that you coached me through removing IPv6 support from my kernel to make sure it was still harmless.
[05:13] <lamont> ah, right
[05:13] <lamont> so specifying ::1 when the system had no ipv6 just meant that it ignored it, yes?
[05:13] <ScottK> Yes
[05:13] <ScottK> It just never matches anything
[05:17] <lamont> sadly, that breaks all the po-files.   Time to find all the other templates changes so we  can have a more complete time.
[05:17] <ScottK> Ahhhh .... how does that happen?
[05:19] <ScottK> lamont: That sounds painful then.  I'd guess we want to do that one a little later then.
[05:48] <soulc> :wq
[05:48] <soulc> er
[05:48] <soulc> ha ha ha
[05:48] <lamont> ScottK: 433660 is more one of me just not wanting to deal with complicating things...
[05:48] <lamont> how common is multiple instances, I wonder?
[05:49] <ScottK> It's extremely useful for complex setups.
[05:50] <ScottK> I don't think there's much you can do with it that you can't do with shoving stuff around in main.cf and master,cf, but eventually complexit there will get you.
[05:50] <ScottK> Multiple instances is, I think, something you want to wait for Weitse to be happy with.
[05:51] <ScottK> Victor Duchovni has had a multiple instance patch that he was working to get into 2.4 and 2.5.  It didn't get in.  I have to assume there's a reason.
[05:51] <ScottK> lamont: ^^^
[05:52] <lamont> oh, most certainly
[05:58] <lamont> ScottK: 153706... I'm wondering if maybe the right answer there is to add to the manpage that queue ids are case sensitive...  mebbe I'll just ask upstream about that...
[05:59] <ScottK> lamont: I think logging failure to find anything is a reasonable request.
[06:00] <lamont> yeah.  likewise, knowing that queueids could be case-insensitive would allow us to DTRT
[06:01] <ScottK> lamont: True, it wouldn't hurt to add it to the man page, but I think logging the failure to find the specified queue id is more important.
[06:01] <ScottK> I'd go for both.
[06:04] <ScottK> lamont: I need to get to bed.  Please shoot me /msg or email if you want me to take action on my bug triaging suggestions.
[06:04] <ScottK> lamont: Good night and good luck.
[06:04] <lamont> ok
[06:04] <lamont> thanks, btw
[06:05] <ScottK> You're welcome.
[07:07] <Iulian> G'morning
[07:36] <soulc> so does cgi-bin install with the apache server?
[11:35] <kraut> moin
[12:43] <sweeties> Hello
[12:44] <sweeties> I am trying to upgrade from 6.10 to 7.04 but the docs say to install update-manager-core, but I can't find that package
[12:45] <sweeties> ?
[12:52] <sweeties> I've checked my repositories
[12:52] <sweeties> and aside from the fact that they are pointing to canadian mirrors, I don't see anything bad
[13:10] <Iulian> Blah
[14:32] <CaptObvious> any ideas how I'd get vesafb-tng going?  is it a module that's included in the server distro by default?
[14:39] <soso__fat> 请教，U7.1服务器 ，如何安装raid5上亚
[14:39] <soso__fat> HP350
[14:39] <soso__fat> 陈列卡，是HP5XX的
[14:39] <soso__fat> 04年的卡
[14:40] <CaptObvious> soso__fat: you're unlikely to get chinese speakers in here
[14:43] <soso__fat> my englishi ,is very very poor
[14:44] <Iulian> soso__fat: You should try.
[14:45] <Iulian> translate.google.com is awesome, afaik
[14:45] <CaptObvious> not sure it does chinese
[14:46] <Iulian> CaptObvious: There are a lot of languages, including Chinese
[14:46] <CaptObvious> cool, didn't know that
[14:50] <CaptObvious> Iulian: any ideas how I'd get vesafb-tng working in the server distro?
[14:55] <soso__fat> how install ubuntu-server7 on raid 5?
[14:55] <soso__fat> I search google,but .......
[15:34] <soulc> yo all
[15:36] <soulc> I have a 22" wide screen monitor and would like to set my terminal to a higher res I think that the default vid driver is limited I have tried editing the grub menu to add vga=792 or 791 but no dice
[15:36] <mok0> soso__fat: I am not sure that the installer supports raid
[15:37] <mok0> soso__fat: but you can try the alternate CD
[15:38] <soulc> how do I install a better vid driver don't really want all of the xwindows running so I am limited to the command line
[15:39] <soulc> which is fine I just don't know how to do these things in ubuntu as I used to run fc7
[15:39] <mok0> soulc: you might try dpkg --configure xorg
[15:39] <soulc> which of course installed and ran EVERTHING
[15:39] <soulc> er
[15:39] <soso__fat> thanks
[15:40] <mok0> soso__fat: come back and ask again if it doesn't work
[15:40] <soulc> if that installs xwindows I really don't want it
[15:40] <mok0> soulc: You don't want x-windows?
[15:40] <soulc> nope
[15:41] <soulc> I would prefer not to have all that running
[15:41] <mok0> soulc: you just want a higher resolution on the console?
[15:41] <soulc> yes
[15:41] <mok0> Hm
[15:41] <soulc> and for some reason vga=792 doesn't work
[15:41] <mok0> soulc: it can be done, I don't know how
[15:42] <soso__fat> ok
[15:42] <soulc> thanks
[15:43] <mok0> soulc: Did any of the other resolutions work?
[15:43] <soulc> well I can change that later as I am ssh'd from my desktop
[15:44] <soulc> how about the cgi-bin dir
[15:44] <soulc> know where or if it was installed?
[15:44] <mok0> soulc: perhaps your monitor doesn't want to switch to that resolution
[15:44] <soulc> did it before in fc*
[15:44] <soulc> I think it is the driver
[15:45] <mok0> soulc: cgi-bin is in /usr/lib
[15:45] <soulc> ah
[15:45] <soulc> thanks
[15:45] <mok0> dpkg -S cgi-bin
[15:46] <soulc> and what does that do?
[15:46] <mok0> soulc: it lists all packages that contain something with cgi-bin
[15:46] <soulc> ah
[15:47] <mok0> soulc: somewhat like rpm -qf :-)
[15:47] <soulc> ah
[15:47] <soulc> ok
[15:47] <soulc> damn I have a lot to learn
[15:47] <mok0> soulc: except it is a pattern search
[15:47] <mok0> soulc: not really
[15:48] <mok0> I switched from centos w/o problems
[15:48] <soulc> well I need to learn how to configure postfix, and where config files are
[15:48] <mok0>  /etc/mail
[15:48] <soulc> like httpd
[15:48] <mok0> yep
[15:48] <mok0>  /etc/apache2
[15:49] <soulc> thanks
[15:49] <mok0> np
[15:49] <faulkes-> soulc: there are several good guides on help.ubuntu.com for setting up postfix that will help yo familiarize yourself with it
[15:50] <soulc> ok good deal
[15:52] <soulc> so is the web server set up as a virtualhost?
[15:54] <soulc> is that site help.ubuntu.com have server specific documentation?
[15:56] <soulc> I wish there was a standard naming convention for apache etc
[15:58] <faulkes-> yes, there is server specific documentation there
[15:59] <faulkes-> iirc, help.ubuntu.com/community/Servers
[15:59] <faulkes-> is community howoto information
[15:59] <faulkes-> you can also find the official server guide there as well
[16:00] <soulc> thanks
[16:20] <ScottK> mok0 and soulc: In Debian/Ubuntu postfix config files are in /etc/postfix.
[16:20] <mok0> ScottK: You're right, sorry soulc
[16:22] <mok0> Just goes to show the deterioration of my memory span. I was fooling around with postfix configuration a week ago
[16:22] <ScottK> mok0: Do you have updated torque packages yet?
[16:22] <ScottK> (from my last round of comments)
[16:22] <mok0> ScottK: no, not yet
[16:23] <mok0> ScottK: there's an awful lot of linitan warnings all of a sudden
[16:23] <ScottK> Are we going to?  Not much time left.
[16:23] <ScottK> mok0: Newer lintian is more pedantic probably.
[16:23] <mok0> ScottK: I'd like that
[16:24] <ScottK> If you want to pastebin me the list of ones you're still sorting out, I'd let you know which ones I think we can live with for now.
[16:24] <mok0> ScottK: There were some warnings about conffiles that I need to work out
[16:24] <mok0> ScottK: ... and I have to implement a chrpath cleaning of the binaries
[16:26] <mok0> http://pastebin.ubuntu.com/4378/
[16:26]  * ScottK looks
[16:27] <ScottK> debian-changelog-file-is-a-symlink can be ignored.  That's due to Ubuntu specific changes to CDBS to save space on the CD.
[16:28] <mok0> ScottK: ok
[16:28] <mok0> ScottK: I hope we can skip the manpage warnings
[16:29] <mok0> ScottK: the ones that are missing a manpage are not really used, perhaps for testing and stuff
[16:29] <ScottK> You should fix the formatting errors.
[16:29] <zobbo> if there a mailing list or rss feed where I can easily see updates that I'll get when I'm doing an upgrade for 7.10 Server ?
[16:30] <ScottK> The conffile ones must be fixed before the first upload
[16:30] <ScottK> zobbo: http://media.ubuntu-nl.org/rss/hardy.xml
[16:30] <mok0> ScottK: yes, they are actual errors
[16:31] <mok0> s/hardy/gutsy
[16:31] <zobbo> ScottK: many thanks
[16:31] <ScottK> mok0: setuid-binary is on purpose I assume?  If so, add an over-ride for it.
[16:31] <zobbo> mok0: I guessed ;) - but ta
[16:32] <mok0> ScottK: I have done that, but not tested yet
[16:32] <ScottK> OK
[16:32] <mok0> ScottK: (it needs to be suid)
[16:32] <ScottK> OK.  Then add an over-ride
[16:32] <ScottK> rpath I'm really not sure about.
[16:33] <ScottK> If you can add the over-ride, fix the conffile stuff, and the man page typos, I think it's worth looking at uploading.
[16:33] <mok0> ScottK: well, they're all pointing to /usr/lib, so it shouldn't matter. But they are not allowed cf. policy, afaics
[16:33] <ScottK> So I don't know that it blocks an initial upload, but could be bug fixed after.
[16:33] <mok0> ScottK: Great, I will see if I can get it done after dinner
[16:35] <mok0> ScottK: I'm glad the courier merge went well
[16:35] <ScottK> It was much better when I applied the patch to the right version ...
[16:37] <mok0> ScottK: Fortunately (for me) I checked that it worked before uploading. But suddenly I feared I made the patch against the wrong sources
[16:37] <ScottK> Which is always a good practice.
[16:37] <mok0> ScottK: I could not use debdiff, because dpkg-buildpackage does strange things to the *.po files (?)
[16:37] <ScottK> I've been doing more Main work recently and I've had to start making debdiffs again as a result.
[16:38] <ScottK> mok0: Yes.  It's an issue that really deserves investigation, but I've never have time.
[16:39] <mok0> ScottK: Just as long as the LoCo teams submit their updates to Debian it will be ok to ignore them :-)
[16:40] <ScottK> Unfortunately Debian and Ubuntu have very different translation workflows and it's very hard to give updated translations back.
[16:40] <mok0> ScottK: Oh, I see...
[16:41] <ScottK> We don't use the Debian tools at all.  It's all magic in Rosetta here.
[16:42] <mok0> ScottK: I tried to copy the po files from debian to the ubuntu tree, build packages from both trees and suddenly there were diffs in the po files
[16:42] <ScottK> Which does point at dpkg-buildpackage.
[16:42] <mok0> ScottK: Indeed
[16:44] <lamont> the lintian bitching about libraries is unsurprising... personally, I'm waiting for the lintian-library police to tell me how to fix my package.
[16:44] <lamont> since it's not immediately obvious
[16:45] <ScottK> ;-)
[16:45] <mok0> lamont: what's the bitching about?
[16:46] <lamont> somethign of the form "links with library that it doesn't use"
[16:46] <ScottK> Yeah.  I've been ignoring those too.
[16:47] <mok0> lamont: I've seen that.
[16:47] <ScottK> Though to be a bit pedantic the blame there lies with dpkg-shlibdeps and not lintian.
[16:48] <mok0> lamont: It's probably because the shared library is linked to libm, say. That you can do as a convinience so people don't need to know all the dependencies of the library
[16:48] <ScottK> If we're gonna bitch, we ought to have the right targe.
[16:48] <lamont> ScottK: it's not lintian.  it's generally that you link with a library that links with some other library
[16:49] <ScottK> Ah.
[16:49] <lamont> which was the result of the _previous_ round of library bitching (not linking with libs that you use)
[16:49] <mok0> lamont: perhaps you should try updating your version of libtool
[16:49] <ScottK> lamont: Are you going to upload 2.5.1~rc1 to Ubuntu?
[16:49] <lamont> mok0: you're presupposing that my package uses libtool
[16:49] <mok0> lamont: yes, isn't it?
[16:49] <lamont> ScottK: yeah - I plan to sync it after today's dinstall
[16:49] <ScottK> Great.
[16:49] <lamont> mok0: no
[16:50] <mok0> lamont: ugh
[16:50] <lamont> if upstream uses libtool, I don't revert it.  if upstream doesn't, I stay as far from that festering pile of garbage as I can
[16:50] <lamont> which is pretty much the same way I deal with auto-crap
[16:50] <mok0> lamont: harsh words on a saturday...
[16:50] <lamont> mok0: well, maybe I'm just old and crufty and not fully awake yet...
[16:51] <lamont> afterall, it's not even 10AM here.
[16:51] <mok0> lamont: :-)
[16:51] <mok0> lamont: 18:00 here...
[16:53] <lamont> ScottK: btw, logcheck winds up way out-of-date with 2.5.
[16:53] <lamont> hrm.. come to think of it, most of my packages are old and crufty too.
[16:54] <faulkes-> don't be hating on us old people
[16:54] <faulkes-> we'll make you get off our lawns.
[16:54] <lamont> LOPL
[16:54] <lamont> LOL, too
[16:55]  * faulkes- wanders off muttering about "damn punk kids" and "newfangled internets"
[16:55] <ScottK> lamont: Well since it's a native Debian thing, they getter get moving then.
[16:56] <lamont> ScottK: well, there's also this longstanding agreement with the logcheck guys that some day postfix will start delivering those files... maybe this is a good point to start that
[17:48] <faulkes-> quick question
[17:49] <faulkes-> n/m, answered it myself
[17:49] <ScottK> That was quick.
[17:50] <zobbo> it is possible to easily determine (or even not easily I guess, but must be scriptable) the last date an update and upgrade was successfully completed ?
[17:50] <zobbo> ignore that
[17:50] <zobbo> let me google for a bit
[17:51] <ScottK> The answer to your question is yes btw.
[17:51] <zobbo> :)
[17:58] <zobbo> hmmm
[18:01] <faulkes-> ScottK: occassionally my brain works in fast mode, not often mind you
[18:01] <ScottK> ;-)
[18:01] <faulkes-> guess the gerbil drank his caffeine this morning
[18:06] <zobbo> hmmm ... well - the best I've found so far is /var/log/apt/term.log ... but that's basically a full text log and I'd have to search through it to see if an upgrade occurred
[18:12] <faulkes-> zobbo: if the file is only written too when an update occurs, then you only need to look at the time the file was modified
[18:13] <faulkes->  /var/log/dpkg.log might be a better option
[18:16] <zobbo> faulkes-: thanks for the help. Problem I can see is that if I install an app as a one off (e.g. i just installed vim), that log file is updated then as well, but obviously the system is not upgraded
[18:17] <zobbo> I could just run a cron script every hour to grep through that and set a time somewhere, if it sees an upgrade happening
[18:17] <zobbo> or I could have a beer I guess
[18:18] <faulkes-> true, and my many years of experience would suggest, the appropriate course of action, would be to start drinking
[18:18] <ScottK> zobbo: What problem are you trying to solve?
[18:19] <faulkes-> <zobbo> it is possible to easily determine (or even not easily I guess, but must be scriptable) the last date an
[18:19] <faulkes-> update and upgrade was successfully completed ?
[18:19] <zobbo> I've got 170 boxes and I want to be able to see from a central server the last time an "apt-get upgrade" was performed on them
[18:19] <zobbo> I can handle the central server bit if I can find out how to get the info from each box
[18:21] <ScottK> zobbo: /var/log/auth* files ought to have that because you need elevated permissions to execute the command.
[18:24] <zobbo> so grep through that (or dpkg) and look for upgrades and write that info elsewhere ?
[18:24] <zobbo> I guess I could tie that in with zabbix
[18:25] <ScottK> That or have the script mail the results to the central server and parse the message there.
[18:26] <zobbo> ScottK/faulkes thanks for the thoughts and help
[18:26] <zobbo> I'll let you know how I get on (in about 3 months time at this rate ;) )
[18:26] <ScottK> zobbo: You're welcome.
[18:40] <faulkes-> hmm, coffee sounds good about now
[18:59] <jondowd> hello. after installing 6.06 lts, what is the terminal or cli command to perform security updates? thanks.
[19:00] <ScottK> jondowd: sudo apt-get update
[19:00] <ScottK> jondowd: sudo apt-get upgrade
[19:04] <jondowd> thanks ScottK - will those commands bring more packages than are needed for security reasons? is there a way to get just the 'security' updates (if there is such a distinction)?
[19:05] <ScottK> That will bring both security updates and stable release bugfixes.  It will not bring general package upates
[19:06] <ScottK> There is a distinction and you can edit /etc/apt/sources.list to not have the updates repository enabled, but I strongly suggest you don't do that.
[19:06] <ScottK> stable updates are tested carefully before upload.
[19:06] <ScottK> Additionally, it does happen that security issues get swept into update packages.
[19:08] <jondowd> ok. I'll do what you recommend - thanks for your help
[19:11] <jondowd> one problem i want to avoid is that in 2.6.22-14-server i get occaisional network failure, so I'm down-grading to 6.06
[19:13] <jondowd> as in "Bug description [edit]
[19:13] <jondowd> Binary package hint: linux-source-2.6.20
[19:13] <jondowd> With linux 2.6.19 and 2.6.20 the network breaks every now and then (might be after some minutes or hours) and the /var/log/messages log mentions "NETDEV WATCHDOG: eth1: transmit timed out".
[19:13] <jondowd> From /var/log/syslog:
[19:13] <jondowd> "...
[19:13] <jondowd> NETDEV WATCHDOG: eth1: transmit timed out
[19:13] <jondowd> eth1: Tx timed out, lost interrupt? TSR=0x3, ISR=0x3, t=356."
[19:14] <faulkes-> the issue could very much depend on the card driver
[19:15] <faulkes-> do you know what card is in eth1?
[19:18] <ScottK> jondowd: You have to do a lot more than apt-get upgrade to get to a new release.  Don't worry about that.
[19:22] <jondowd> per-fect ! (my download of ubuntu-6.06.2-server-amd64.iso is at about 94%) thanks again!
[22:02] <[diablo]> evening guys
[22:02] <[diablo]> does anyone know if the 6.06.2 release fixes the gfxboot issue for KVM guests please?
[22:03] <[diablo]> found that soooo frustrating
[22:06] <ivoks> no
[22:06] <ivoks> but that's easily avoidable
[22:06] <[diablo]> hi ivoks
[22:06] <[diablo]> ivoks, yeah I know... just remaster the cd
[22:06] <ivoks> or use qemu
[22:06] <[diablo]> pfff
[22:06] <[diablo]> blow that
[22:07] <[diablo]> the installation takes for ever
[22:07] <[diablo]> remastering is quicker
[22:07] <ivoks> well, with kqemu, should be enough for install
[22:07] <[diablo]> nod
[22:07] <[diablo]> that's true, but I don't run kqemu
[22:07] <[diablo]> just kvm
[22:08] <[diablo]> actually I personally feel that this gfxboot issue should have been fixed for 6.06.2
[22:08] <[diablo]> would have been so simple
[22:09] <ivoks> it's not gfxboot issue
[22:09] <ivoks> but a kvm issue
[22:09] <[diablo]> no
[22:09] <[diablo]> well
[22:09] <[diablo]> erm
[22:09] <[diablo]> a bit of both
[22:10] <[diablo]> I know it's been fixed in 8.04 alpha's
[22:10] <[diablo]> actually it works great in a kvm-60 guest
[22:24] <hsn_> !db2
[22:24] <ubotu> Sorry, I don't know anything about db2 - try searching on http://ubotu.ubuntu-nl.org/factoids.cgi
[22:24] <hsn_> !package db2exc
[22:24] <ubotu> Sorry, I don't know anything about package db2exc - try searching on http://ubotu.ubuntu-nl.org/factoids.cgi
[22:25] <ScottK> hsn_: It's in the partner repository
[22:26] <hsn_> how to add it?
[22:26] <ScottK> edit /etc/apt/sources.list  It'll be called commercial or partner depending on which release you are on.
[22:34] <hsn_> ah found it
[23:54] <[diablo]> btw, 6.06.2 DOES fix the kvm gfxboot issue
[23:54] <[diablo]> !