[00:16] New bug: #264155 in mysql-dfsg-5.0 (main) "fails "subselect" on i386 testsuite runs with segv when compiled with PIE" [Undecided,New] https://launchpad.net/bugs/264155 [00:51] hi room [00:51] will a dell dimension 5150 dual core 145gig harddrive make a good server [00:53] For what? [00:57] I want to make a server for my family who are out of state to access for picture sharing [00:58] this 10meg email limit is killing me [00:58] How much RAM for this beast? [00:59] well it can hold 4gigs [00:59] I was going to put 2 500gig hard drives in it [00:59] i think what you need is a p2 433, with 256mb of ram, and adsl2 or greater [01:00] is that all [01:00] will that hold ubuntu server [01:00] bump it to a 667 with 512mb of ram if your hosting images+thumbnails in a website [01:00] and run it [01:00] wobbiebobbie: So can I, but seriously, you need to think about what use it's going to be put to. If you're going to run 17 zillion web-apps, then you need a big machine, but if it's going to be doing bugger all, then it can be the size kgoetz suggests. [01:01] I just like to set it up to share pictures and file with my family though a FTP [01:02] let ftp die. use some other protocol [01:02] like what I dont know [01:03] I have ubuntu gusty but have not used ubuntu server [01:03] http at a pinch, ssh if posible, [01:03] http wouldnt be such a bad idea, imho. [01:04] is there a how to for ubuntu server [01:04] theres a server guide on help.ubuntu.com [01:05] thanks I will check it out [01:05] heck I was going to buy a server from dell thanks guys [01:06] heh. np. feel free to ask if you need a hand (afer checking th docs :P) [01:06] ok thanks [01:08] kgoetz: You know we're going to get more questions about setting up a server to share photos with the family don't you :) [01:08] owh: how hard can it be? intall ubuntu server -> use lamp task -> log in to server -> mkdir public_html -> photos in public_html -> ??? -> profit [01:09] ROTFL [01:09] kgoetz: So how's sunny ADE? [01:09] * owh hasn't opened the curtains here yet. [01:09] owh: sunny, for two days in a row [01:09] Whoa, there's sun here too! [01:09] :O [01:09] quick, hide before it sees you! [01:10] * owh blinks in the bright light. [01:11] * kgoetz is at work battling php/mediawiki/ldap (and all that it entails) [01:12] Oooh, let me know what you think about mediawiki when you're done. I'm looking at using it for a project or six. [01:12] making it use htauth is easy (then use apache to backend to ldap), but using mediawikis ldap plugins ... not so easy. i expect today is a day of doco reading and code poking for me [01:13] other then the php thing, i quite like mediawiki [01:20] kgoetz: I've been playing with PHP for longer than I care to admit. I think it was still called FI, or some-such, PHP2. [01:20] hehe [02:05] if i add a new SATA device to a live system is there a way I can make the kernel/driver rescan the bus? [02:43] docta_v: Under dapper I played with rescan_scsi_bus or some such script, but it was never very successful. [02:50] owh: cool.. i don't think it's going to work on this system [02:51] in dmesg it says it shutdown the port [02:51] i'm emailing the devs and telling them to get off the box tomorrow at lunch so i can take it down [02:59] docta_v: Really it should be plug and detect and I've been stumped on how to achieve that. [03:11] owh: well you need the right controller in addition to the right kernel, etc. [03:11] my controller is silicon images [03:12] and all of their hardware is garbage so i'm assuming that's why it doesn't work [03:12] :) [03:14] i think you can do scsiadd -s [03:14] provided your controller handles the hot-plug [03:33] docta_v: I recall that command, but not the outcome, other than that it didn't work at the time. Something to do with a timeout IIRC, but that might be a reflection on poor hardware as you already commented on. === freaky[t] is now known as fReAkY[t] [03:47] Hey guys I'm having an issue with installing x11 on an old machine running server 8.04 [03:47] I get the error : Framebuffer bpp 32 not supported for this chipset === fReAkY[t] is now known as freaky[t] [03:51] los invitamos a #linux-para-todos >>> Un canal libre para la Gente Libre, donde tratamos de ayudar sin tantas reglas, donde esta permitido expresarse, copiar y pegar, y ser un poco mas abierto. donde no importa un pais, sino la libertad de usar Linux. [03:53] maestrolinux: English around here... [04:44] a free route for free people, where we can try to help without as much regulation, where one is permitted to express, copy and paste, and be a little more open, where country is not important, "sino" the freedom to use Linux. [04:45] sino, but rather [04:45] ah [04:46] So, I had this idea, about having a setup in tasksel to let people choose what scripting backend to have on their LAMP: PHP, Python or Ruby(oR), with the click of a checkbox. Is something similar to this being worked on, or should I formally bring this up somewhere? [04:47] missing the other P (Perl) [04:47] ya, sry, I'm not a perl guy [04:48] but yes [05:12] lukehasnoname: That's an excellent idea. [05:12] lukehasnoname: Send an email to the list and ask for some other opinion would be my first recommendation. [08:06] moin [08:07] moin [08:27] What is "Framebuffer bpp"? [08:27] <_ruben> bits per pixel i'd guess [08:27] <_ruben> and thus the color depth of the pixels of your framebuffer/console [08:28] Ah k thank you [09:30] Hi., I have problem about mail server ., Can you help me?. Please [09:31] <_ruben> no, we cant, since we dont know what the problem is [09:32] Ohh..... [09:32] I use courier-maildrop from apt-get. and i create user vmail for maildrop [09:34] when i use maildrop from root accout it say " maildrop: Changing to /home/vmail/npt.com" and it right! [09:35] but when i use maildrop from vmail account, it say "maildrop: Changing to /home/vmail" and i wrong [09:36] i am forgetful , i use courier-authlib-ldap. and i get user account from ldap. [09:36] :-( [09:36] so i try change vmail 's home directory [09:36] export HOME = / [09:37] and use maildrop again, it say "maildrop: Changing to /" [09:38] i don't know where is make problem [09:57] - - [09:58] sorry, I've only used dovecot & exim for mail, I can't help [10:00] thank you , no at all :-) [10:11] Even though this is a more general ubuntu question. How do I get the "add packages" button into the Applications menu after I installed gnome on ubuntu server? [10:12] elnewb: I'd guess 'sudo apt-get install gnome-app-install' [10:12] Seems to be working away thank you [10:13] Works [10:29] looking into monit, what other similar packages are out there please? [10:36] hi [10:39] how do i use rsnapshot on ubuntu? i need a workaround for that .gvfs fuse crap :( [10:39] using "exclude" in rsnapshot.conf does not work -- rsync still readlink()s the stuff and errors... [10:44] ... so you are basically forced to ignore all errors. crap crap crap. [10:45] moldy: that's sad to hear. I use it on an older system, and will be upgrading to hardy soon [10:46] then you're in for some trouble [10:46] i now found out: using "exclude .gvfs" rather than "exclude .gvfs/" works around the problem [10:46] can you just turn off gvfs? [10:46] not easily, afaik [10:47] root not being allowed to read stuff ist just plain idiotic... [10:47] what are the fuse guys thinking... [10:47] fuse guys or nautilus guys? [10:47] afaik, it's a fuse thing [10:47] fuse mounts are only readable by the owner, not by root. idiotic. [10:51] To lift this restriction for all users or for just root, mount the filesystem with the "-oallow_other" or "-oallow_root" mount option, respectively. Non-root users can only use these mount options if "user_allow_other" is specified in /etc/fuse.conf. [10:51] moldy: found above on fuse wiki. But I only understand half of it [10:53] jimcooncat: hm, i think i understand it [10:53] what i don't know is where/how gnome issues the mount command [10:53] and where you can manipulate it to include -oallow_root [10:54] there also is a bug in the ubuntu bts about this [10:54] moldy: maybe you could manually mount it? [10:54] it's somehow magically mounted upon login or something like that [10:55] using "exclude .gvfs" is the best workaround i found so far === mdz_ is now known as mdz [11:01] i can see why the fuse guys are doing this for *some* cases (remote mounts) [11:01] but in the usual case it's just idiotic... the default should be changed [12:18] morning all [12:19] hi mate [12:20] hello [12:29] kirkland: if you have some time, could you please review the patch that we have in cryptsetup for the additional 'waiting-for-root' loop? http://paste.ubuntu.com/43031/ [12:29] kirkland: https://bugs.edge.launchpad.net/ubuntu/+source/cryptsetup/+bug/251164/comments/6 claims that the current approach will break the failure hooks [12:29] Launchpad bug 251164 in cryptsetup "crypdisk boot scripts do not wait properly for source devices to appear" [Undecided,New] [12:30] kirkland: I also think that the current patch is not optimal, so I'd like to have some expert's opinion on this [12:41] New bug: #264306 in dovecot (main) "Panic: file sieve-cmu.c: line 88 (unfold_header): assertion failed: (str[i] == ' ' || str[i] == '\t')" [Undecided,New] https://launchpad.net/bugs/264306 === freaky[t] is now known as fReAkY[t] === fReAkY[t] is now known as freaky[t] [12:59] morning [13:13] morning [13:13] should this be present in a users directory "/home/karla/ /crond/192.168.1.79.user2:" [13:14] I had a virus infect this particular user email box called [13:14] Linux/Rst.a === freaky[t] is now known as fReAkY[t] [13:53] siretart: I'm happy to review the patch, but you're generous in calling me an "expert" ;-) [14:41] zul: are you asking for my dovecot.conf and sieve.conf because you believe they are necessary to diagnose the bug, or just as a matter of course? [14:42] zul: because I think I understand the problem already and it has nothing to do with my configuration [14:42] mdz: thinking about it now just the sieve.conf [14:44] kirkland: :-) [14:44] siretart: i'm reviewing it now [14:45] zul: you can reproduce it with a one-line sieve.conf as follows: [14:45] zul: if header :contains "Subject" "blahblahblah" { discard; } [14:45] kirkland: in case it's useful for you, you might want to checkout the bzr branch on launchpad. it has the debian source imported in its history, so you can compare files against the version in debian like you can see in the paste [14:45] mdz: thanks [14:45] kirkland: thanks! [14:45] mdz: will try to get it fixed today [14:45] siretart: can you give me the lp URL? [14:45] kirkland: apt-cache showsrc cryptsetup ;) [14:46] Vcs-Bzr: https://code.launchpad.net/~ubuntu-core-dev/cryptsetup/ubuntu [14:46] zul: I believe the problem is that the subject contains MIME-encoded newlines [14:46] Subject: =?windows-1255?Q?=E0=E9=EE=E5=EF_=E0=E9=F9=E9_=EC=E9=F8=E9=E3=E4_=E1=E8=E5=E7=E4_=E1=EE=F9=F7=EC=0D=0A?= [14:46] or 'bzr checkout lp:~ubuntu-core-dev/cryptsetup/ubuntu', if you intend to commit [14:46] note =0D=0A [14:46] mdz: there might be a fix in the dovecot mercurcial repos already I just have to test it out === fReAkY[t] is now known as freaky[t] [14:58] siretart: I'm inclined to agree with ceg, though, that cryptroot should be able use the failurehooks like we recently did for md [15:00] siretart: the panic -r call is what starts the failure hooks [15:04] kirkland: I'm not familiar at all with the failurehooks. Unless you can point me to documentation how it is supposed to work, could you perhaps give the cryptroot hook a shot to make it work gracefully with the failurehooks? [15:04] siretart: that's what I'm looking at now ;-) [15:06] :-) [15:06] siretart: for documentation, see the comment in /usr/share/initramfs-tools/scripts/functions:try_failure_hooks() [15:07] siretart: (assuming you're on an up-to-date Intrepid system) [15:07] siretart: otherwise, initramfs-tools-0.92bubuntu10/scripts/functions [15:08] aah, I see [15:08] siretart: so you'd put some magic in /tmp/mountroot-fail-hooks.d/cryptroot [15:09] where is the doc that tells me how to upgrade from 6.04lts to hardy lts? [15:10] for server -- i cant seem to put my keys on it today [15:15] siretart: okay, and for your example, apt-get source mdadm; vi mdadm-2.6.7/debian/initramfs/init-premount [15:15] siretart: you'll see at the bottom of that script it "installs itself" with add_mountroot_fail_hook "10-mdadm" [15:16] siretart: and note that it exits "0" when it thinks that it has done some good, "1" otherwise [15:43] lamont: I see in debian/changelog you added my filter/policy server scripts to postfix, but I don't actually find them when I inspect the .deb? [15:44] And with that, I run out the door for several hours. [15:44] meh [15:45] dpkg -L postfix| grep add [15:45] /usr/sbin/postfix-add-policy [15:45] seems there for me.... [15:47] heh. OTOH, I guess I should deliver manpages, too. :-( [15:47] in section 8, not 1 [16:19] kirkland: I don't know if I'll have time to test and implement the failure hooks myself for cryptsetup. if you have some spare time, I'd really appreciate it if you would do it [16:19] siretart: I'll see what I can do...... [16:19] siretart: can you detail how to reproduce the failure case? [16:19] siretart: such that I can test the work I would theoretically do for you? [16:22] kirkland: AFAIUI boot from root on usb would trigger it. I never experienced that personally yet. [16:24] siretart: boot from crypted root on usb, you mean? [16:25] yes [16:27] siretart: you think usb flash disk is sufficient? [16:27] siretart: i have a few of those lying around ;-) [16:28] kirkland: if it has an on board ide/sata controller, then probably not. a normal usb stick should do it however [16:28] kirkland: a really slow scsi control should trigger it as well, though [16:28] the point of the excercise is to have a really slow controller for the root device [16:28] siretart: sorry, yeah, i mean normal usb stick [16:30] yes, something like that should trigger it. it depends a bit on the usb controller, but usually, they take a loooooong time to initialize [17:01] New bug: #264374 in bacula (universe) "package bacula-director-pgsql 2.2.8-5ubuntu7 failed to install/upgrade: subprocess post-installation script returned error exit status 1" [Undecided,New] https://launchpad.net/bugs/264374 [17:06] that bug is now owned by carl [17:09] can somebody tell me how east it is to setup software raid on ubuntu server? [17:10] Quite. [17:11] soren: how quite? [17:11] I was going to say it depends on where you live. [17:14] wayneandleanne: Depends. Do you have a unit of measurement? [17:15] wayneandleanne: It's really quite easy. [17:16] Is it 'open a jar of pickles' easy, or 'installing KVM' easy? [17:16] ok, i have a machine i want to run ubuntu server on (dns, samba, asterisk,etc) for home use in has a 40 gb drive i am going to install server onto and a pair of 80's i want to mirror but i dont know how to [17:18] In the installer, for each partition you want to mirror, you create a partition on each disk. [17:18] You mark this as a "raid member". [17:18] Once you've done that you choose to "configure software raid". [17:18] In there, you create the raid devices and add your filesystems to that. [17:18] That's about it. [17:18] that easy? [17:19] thanks [17:19] np [17:23] soren: is it basically the same steps to add extra hd's at a later date? [17:33] hi there! [17:34] yo [18:03] nijaba: this may scare some people off: "The record kept of your survey responses does not contain any identifying information about you unless a specific question in the survey has asked for this." do we want to clarify that (as I assume) such revealing responses will only be used for aggregate statistics and not be combined with other responses in ways that would compromise privacy? [18:05] nealmcb: did you see the link to the privacy policy? [18:05] yes - but many may ignore that... [18:07] nealmcb: well, if they are scared and don't read the privacy policy, I guess they won't take the survey and that's it.... [18:11] New bug: #264418 in samba (main) "net usershare add : read only (post 43674)" [Undecided,New] https://launchpad.net/bugs/264418 [18:11] nealmcb: the truth is, I am not sure how to prevent this text from appearing [18:12] nijaba: yeah - gotta warn them. [18:13] If any better ways to defuse that occur to me I'll pass them along.... [18:13] thnaks [18:13] thanks too [18:13] * nealmcb wonders if thnaks taste good [18:14] must be a snack with a french accent [18:16] ScottK: is someone in pkg-debian working on this : http://paste.ubuntu.com/43105/ ?? [18:17] leonel: Not that I've heard. They are still just getting started, so I think we should proceed as normal (you're welcome to join BTW, I've just not caught up with you to say so). [18:18] ScottK: count me on .. [18:18] ScottK: I'll start checking diffs for those 2 later this day or tomorrow [18:19] Thanks. [18:31] New bug: #264427 in openldap (main) "openldap user not in ssl-cert group" [Undecided,New] https://launchpad.net/bugs/264427 [18:41] ScottK: clamav-0.94 introduced libclamav.so.5 ... [18:51] * delcoyote hi === freaky[t] is now known as fReAkY[t] === fReAkY[t] is now known as freaky[t] [20:00] mathiaz: The server guide is at the same place as the other Ubuntu documentation? [20:00] jpds: yes [20:01] mathiaz: OK; I will have a crack on it tomorrow. [20:01] jpds: we trying to figure out a way to move it around - but it's not done yet [20:01] jpds: so you'll have to checkout all of the documentation which can take some time the first time [20:01] mathiaz: I know, I've done it before. I used to write on the PackagingGuide long ago. [20:06] hello all, I have a problem for two weeks, I installed a pptp vpn server to my company, I use dydns to keep track of the ip, I have three servers www, dns, and samba, the router is the gateway, when my clients connect, they suffer from a very slow connection because they use my dns instead of their own one [20:16] leonel: Yep. We get to rebuild and fix all the rdepends again. [20:27] jdstrand: hi. From your tests, it is not possible to define an Apache and on ApachE profile. But this is possible and can work when you do manual test… (I handled it specifically in my code) [20:34] jdstrand: ok, my bad, I adapted the result file. The only thing is that all profiles in tests/$class/$subclass/orig/* are the same: not symlink but hard copy. Why? (I have to copy my profile changes in every */orig directories) [20:58] didrocks: actually, it should be a symlink to tests/defaults [20:58] didrocks: but it seems bzr isn't doing that right (or I'm not) [20:58] (probably the latter) [20:59] ok, so, for the moment, I will try to retablish this [20:59] didrocks: you are working off the bzr branch, correct? [20:59] I am adding some tests (I will have to stop because of giving a talk in -classroom), but I will keep you in touch [20:59] for the tests? I am working in my branch, why? [21:00] didrocks: I was just trying to figure out if it was 'bzr export' or not [21:00] I'll look into it [21:01] ok, I will keep every changes in a dedicated place [21:30] mathiaz: Are you tracking updates to the Server seed or is that assigned to someone else? [21:48] mathiaz: hey, are you around? === Laleh__ is now known as Laleh [23:04] ScottK: I don't know of anyone assigned to track seed updates for the server team - any core-dev can change the seeds AFAICT. [23:04] kirkland: yop :) [23:05] mathiaz: no worries, i think kees answered my questions, thx!