=== sixstringsg|away is now known as sixstringsg [01:11] I'm looking for a way to be able to reinstall my entire server, with all current packages and config files included, in the case that my primary and backup servers both go down simultaneously. Is there a good way to do this? I tried using Remastersys, but am unable to reinstall the backed-up system to the disk from the command line (maybe I have the wrong commands, but it seems like it requires a GUI to install) [01:55] k3vmcd: save the output of 'dpkg --get-selections' and backup /etc. That should let you reinstall packages and have the config files to go with them. [01:57] ah yes, that makes sense [01:57] thanks greppy [01:58] probably want to snag some other directories/files as well [01:58] like /var/spool/cron [01:58] but that should get you started. [01:59] the thing to do is to TEST your backup/restore procedure before you need it. [01:59] so shove it all back onto a clean install in virtualbox or something to make sure it works. [02:01] i'm thinking if I back up /etc and /var then I'll capture everything i need [02:02] i think my approach with Remastersys was a bit of an overkill [02:02] k3vmcd: /var includes a bunch a log files which you typically don't need [02:03] true, but it does have my web server files and my SQL databases [02:10] k3vmcd: I didn't mean to suggest that you exclude /var, just /var/log, unless you need them. [02:20] how doth I maketh another account easily??? [02:20] how doth I makith another user account easily??? === sixstringsg is now known as sixstringsg|away [02:22] hey, thanks jeffrubic and greppy. I was just looking through my server files after your suggestions and I figured out a good set to keep backed up. Just wanted to say thanks for the help [02:27] Zilly: try the "adduser " command. Alternatively you can search on how to use the "useradd" command instead [02:29] k3vmcd: I've done that and I have the user, it's just when I login with that user I have to reboot [02:29] i have to reboot because nothing is there [02:29] just blankness [02:30] anything I can get that will set up the user completely? [02:31] edit fstab and set everything up in X? [02:31] are you ssh-ing in? [02:32] no, vbox [02:33] hmm, I ran into something similar when the user was set to /bin/false but I think that might only apply to ssh [02:34] Zilly, 'useradd' is the base command and requires options to set a user up, but the 'adduser' command does set everything up, creates the home directory, all that [02:34] does the user have a folder in /home ? [02:35] going to try adduser, I think I already did but am uncertain now [02:35] k3vmcd: yes [02:36] if I'm not back it didn't work [02:36] bbl === sixstringsg|away is now known as sixstringsg [03:27] I guess it didn't work :( === hayer_ is now known as hayer [04:05] hallyn: is lxc within lxc working for you? I just tried to setup some automated d-i testing in an LXC container instead of a VM and the containers won't start in there [04:05] hallyn: with no clear error in the log or dmesg [04:07] hallyn: http://paste.ubuntu.com/838620/ and http://paste.ubuntu.com/838621/ (not sure that really helps ...) [04:08] hallyn: as you can see, I tried a few times ;) the veth pairs get created and never cleaned, the container never starts [04:09] oh, and I just noticed I don't have cgroup-lite (running with --no-install-recommends by default here ...) [04:09] that'd explain it I guess :) [04:09] yep, that was it ... [04:09] hallyn: sorry ;) === sixstringsg is now known as sixstringsg|away [07:51] New bug: #930881 in samba (main) "smbd crashed with SIGABRT in rep_strlcpy()" [Undecided,New] https://launchpad.net/bugs/930881 === mrmist_ is now known as mrmist === Cluster is now known as ^Blackflowers^ === Resistance is now known as EvilResistance === Internaut is now known as Guest48794 [10:00] Has abyibe akready looked at http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=652026 with regard to ubuntu (12.04) I suppose ubuntu is affected as well [10:00] Debian bug 652026 in amavisd-new "perl 5.14 broke use of s-s-d --name on initscript" [Important,Fixed] [10:03] rdjmhgh, #ubuntu+1 [10:03] which i told you earlier :P [10:41] New bug: #930916 in amavisd-new (main) "amavis start-stop script fails to stop amavisd" [Undecided,New] https://launchpad.net/bugs/930916 === gustav is now known as beerbro [11:42] hi [11:42] as someone tested OpenNebula and Eucalyptus ? [11:42] I heard from a friend that OpenNebula is much more stable than Eucalyptus [11:43] but Ubuntu seems to promote Eucalyptus [11:43] any advice about which one I should use ? [11:55] hi [11:55] maxagaz: euca has been stable for us and working well [12:10] maxagaz: any particular issue known or just gut feeling? === MYANMAR60273 is now known as greenhost [12:10] hi all [12:11] Tm_T: a former colleague used euca before, and has now switched to open nebula, he just told be that now he doesn't any problem while he add a lot of little problems with euca [12:11] Tm_T: but no particular issue known [12:12] who have done ubuntu mail server for windows clients? [12:12] Tm_T: and it's neither just a gut feeling [12:14] who have done ubuntu mail server for windows clients? [12:21] maxagaz: I see [12:22] greenhost: I don't think client OS does matter (: [12:23] ah [12:23] was going to give this for greenhost: https://help.ubuntu.com/community/MailServer [12:24] who have done ubuntu mail server for windows clients? [12:25] 1423.30 < Tm_T> was going to give this for greenhost: https://help.ubuntu.com/community/MailServer [12:25] 1422.33 < Tm_T> greenhost: I don't think client OS does matter (: [12:26] as long as the client is using sensible protocol (imap, pop), ofcourse [12:26] it seems to me that dnsutils is a package related to bind9. But if I want to use dnsmasq instead of bind, do I uninstall bind9 and dnsutils, and thus I will no longer have access to dig or nslookup ? [12:26] Tm_T I installed today with postfix and dovecot [12:26] Tm_T . not ok. [12:27] zastaph: there shouldn't be a problem to have both packages installed? [12:27] Tm_T : I frist tried as mentioned on Postbasichowto official documentation. but starting from netcat mail.yourdomain.com. I got error. [12:27] tm_t, well dnsmasq can't start because port 53 is in use with /usr/share/named [12:27] tm_t, but perhaps dnsutils and dnsmasq can be together [12:28] zastaph: I would assume you can disable the other one if you don't need it [12:28] greenhost: can you be more specific what kind of problems you had with, for example, dovecot? [12:28] tm_t, yes im going to uninstall bind9 package, but just wonder if I should get rid of dnsutils too... nslookup and dig seems like useful tools [12:29] zastaph: no need to get rid of dnsutils AFAICS [12:29] zastaph: there's no reason to remove bind9 [12:29] zastaph: just stop the sevice [12:29] service* [12:29] Tm_T I tried /etc/init.d/dovecot restart. but it looked like that file is corrupted [12:29] jacobw, ok.. but I have no need for bind9 if I'm going to use a different dns tool [12:30] Tm_T " process(58888) error. bla bla [12:30] greenhost: what file? and you have to be more specific [12:30] Tm_T as I am now at cyber cafe with winows XP. I dont remeber all my errors and try at the same time [12:30] greenhost: roger, just hard to know what went wrong (: [12:31] Tm_T: actually I would like to use email from any mail clients( thunderbird or outlook2003) or something available. I first just want to email locally.so how coud I configure on postfix and dovecot? [12:32] greenhost: there seems to be rather clear and good howto at https://help.ubuntu.com/community/Dovecot [12:32] zastaph: bind9 has a small footprint [12:32] zastaph: unless the dnsmasq package conflicts, there's no reason to remove it [12:32] jacobw, and whats the reason to keep it? :) [12:33] Tm_T . you are right. but there are more security based complicated packages to install from internet whenever I read documentation. I have no internet at home on my mail server. so i want to be ok with installer CD alone [12:33] zastaph: there's none, i'm just saying that there's no reason to remove a package to install a similar package [12:34] Tm_TI just would like to know what line in postfix and what lines in dovecot to configure or change to email from user1 to user2 locally from windows mail clients [12:34] Tm_T it is pity there is no English version simple video for that to see step by step [12:35] Tm_T even on youtube [12:36] greenhost: there's is nothing special to do. simply mail from user1 to user2 locally should normaly works out of the box [12:38] jojoburk: just testing from CLI: mailing from root to another user called user1. is fine in CLI alone. but to email really from windows using mail clients need extra configuration i think. I was never been successful [12:40] greenhost: take a look at the output from netstat -tulpe to see on which address the dovecot and postfix server are bound. possible just on localhost? [12:42] jojoburk yes just on localhost . last time [12:45] jojoburk: my question is so simple. I wouldl really like to make sure if it is possible to set up mail server for windows clients with server installer CD alone without internet [12:46] greenhost: when postfix and and dovecot are installed, yes it is. [12:47] jojoburk: can I also install postfix and dovecot using server installer CD without internet? [12:50] greenhost: postfix yes, but i'm not sure about dovecot. give it a try: just try installing dovecot using apt-get or aptitude only with the cd as source [12:52] greenhost: otherwise you can download dovecot package from other pc and copy and install it on your server [12:52] jojoburk; honstly I followed https://help.ubuntu.com/community/PostfixBasicSetupHowto, I was stucked since sudo apt-get install courier-pop [12:52] sudo apt-get install courier-imap [12:53] jojoburk: yes. I am on windows XP machine now. there is no linux here at cyber. How could i copy/download required full packages where there might be dependencies [12:53] ? [12:57] greenhost: so you want mail delivery between local users accessible by imap? [12:58] jacobw: imap or pop or whatever. I don't mind. As long as I can email between local users. it is fine for me [12:59] jacobw : but using any email clients from windows [12:59] yes [12:59] it seems that his problem is to install a mailserver without internet connection. i understood that dovecot is not installed at the moment. [13:03] you can use apt on your ubuntu server to find out what packages you need, but its laborious. [13:04] apt-get -s install will list all packages that would be installed if you ran apt-get install [13:04] apart from that, see http://library.linode.com/email/postfix/dovecot-system-users-debian-6-squeeze [13:06] can i install vmware image in kvm ? if yes which file i suppose to import (from which i will able to boot) [13:07] jojoburk: I took my home server to office to download courier-pop and courier-imap to follow the guide line. but not ok [13:08] jojoburk: downloading was ok. Though DNS was fine, I could not proceed [13:08] samba35: to run a kvm/qemu vm with libvirt on ubuntu server you need a qcow2 disk image and an xml definition [13:09] ok [13:10] jojoburk I would like to have somebody to pointout what i might be wrong in which steps based on "https://help.ubuntu.com/community/PostfixBasicSetupHowto", [13:10] samba35: you can convert a vmware disk image (vdi) to qcow2 using qemu-img convert foo.vdi -O qcow bar.img [13:12] ok let me try [13:15] greenhost: i dont can help you at the moment, im sorry [13:15] jojoburk: thanks anyway [13:16] greenhost: that wiki page is outdated [13:16] greenhost: sometimes the community documentation is [13:16] jacobw: where should I see updated one? [13:18] greenhost: dovecot has been the recommended mda at least 8.04 [13:18] jacobw: is it better than courier? [13:18] greenhost: try following the document I linked earlier [13:18] http://library.linode.com/email/postfix/dovecot-system-users-debian-6-squeeze [13:22] hi guys - I'm seriously struggling to get PSAD working on 10.04 (with UFW). I've just installed a new minimal installation on a web server, and it's not logging anything. Previously (on 8.04) I configured sysklogd to write to psadfifo, but that no longer seems to be an easy option with rsyslogd? [13:23] jacobw it looked good and too much configuration. But it used "internet site" settings. If I wanna use, locally only. What else should I change.? [13:24] jacobw: it is pitty I can't follow and install packages without internet [13:26] jacobw: http://pastebin.ubuntu.com/838994/ [13:26] jacobw: http://pastebin.ubuntu.com/838994/ [13:27] tjaalton: hey, what's the current packaging status of freeipa in ubuntu/debian? Anything I could do to help? :) [13:27] greenhost: the joke is that you have to configure postfix for internet site. locally only means that postfix is just configured for the use on localhost! [13:28] kklimonda: if you're an archive admin then sure ;) [13:28] jacobw, there is no .vdi file there are vmx, vmdk,nvram ,vmsd,cmxf file are there [13:28] jojoburk: do you mean I have to configure as internet site though my mail server is not without internet? === ABibhuti- is now known as AAYUSH [13:28] greenhost: so your ubuntu server can only access the LAN? [13:28] tjaalton: it's already that far? :) [13:28] kklimonda: most of it is packaged, some are waiting in NEW [13:28] samba35: yes, i made a mistake, vdi is the format for virtualbox [13:28] pretty far yes [13:28] jojoburk jojoburk: do you mean I have to configure as internet site though my mail server is without internet? [13:29] kklimonda: though I haven't tested server install yet [13:29] where can I download dovecot full tgz package? [13:29] tjaalton: where do you keep all the packaging? I'd give it a shot to see how it breaks ;) [13:30] samba35: vmdk is the format for vmware, qemu-img supports images from vmware 3 and 4 [13:30] ic [13:30] kklimonda: git.debian.org [13:30] greenhost: yes, because "internet site" means that die daemon is accessible from "other" hosts in network (no difference between internet and lan). [13:30] how do i check whether image is 4 or 5 [13:31] jojoburk : That is the very important point. [13:31] greenhost: ... and not just from the server itself (localhost) [13:31] hi guys - I'm seriously struggling to get PSAD working on 10.04 (with UFW). I've just installed a new minimal installation on a web server, and it's not logging anything. Previously (on 8.04) I configured sysklogd to write to psadfifo, but that no longer seems to be an easy option with rsyslogd? [13:31] kklimonda: there is no easy way to test it atm though [13:31] ok file give me results its 4 [13:31] CentOS5.2small.vmdk: VMware4 disk image [13:31] samba35: you might want to try using virt-convert on the directory containing the images and definitions to get a qcow2 image and libvirt xml definition [13:31] jojoburk: I am now leaving cybercafe. so I might need to download dovecot tgz so that I can take it home with flash drive. [13:32] hi fluffypony [13:32] jojoburk where could I get dovecot full package tgz file? [13:32] ok [13:32] hi jacobw [13:32] just search for dovecot download? [13:33] CentOS5.2small.vmdk: VMware4 disk image ? [13:33] sorry [13:33] virt-convert CentOS5.2small.vmdk ? [13:34] don't download the tarball, its a source package not packaged for ubuntu [13:34] building dovecot would not be trivial [13:34] tjaalton: I'm really interested in getting it packaged (and working) in Debian and Ubuntu. I don't really expect it to work out of the box right now, I was just interested where the packaging is so I can start working on it :) [13:34] ok [13:34] kklimonda: https://launchpad.net/~ubuntu-389-directory-server https://launchpad.net/~freeipa [13:35] kklimonda: pki is mostly packaged, but upstream is keeping it all in pki.git, so would need to merge it all back in at some point, maybe not for this push though [13:35] greenhost: run `sudo apt-get -s install dovecot-common` on your ubunu server, note what package it wants to install, and get the DEB files from packages.ubuntu.com [13:35] kklimonda: there are ppa's too, but the oneiric ones are not fully uptodate [13:36] samba35: virt-convert the directory [13:36] best to test stuff against precise at this point [13:36] yeah, I'm already working (and testing stuff against) precise [13:36] kklimonda: are you on collab-maint? [13:36] alioth [13:36] tjaalton: yes, I should have access via pkg-gnome [13:36] ok [13:36] ok [13:37] jojoburk: I want to download from windowsXP. no apt-get. I know I can download from package.ubuntu.com. but nested dependencies. how to take it all? [13:37] fluffypony: what have you done so far? [13:37] kklimonda: i'll probably set up a project around freeipa, like there is pkg-fedora-ds [13:37] same for sssd/ding-libs [13:38] jacobw: psad is installed and setup (same as on 8.04) and configured to my liking. added the two iptables rules straight from the cipherdine site. psad -S status shows no packet counts? [13:38] greenhost: http://wubdepends.sourceforge.net/ [13:39] tjaalton: great to hear that :) [13:40] kklimonda: actually did that too, but the admins asked if it could be under collab-maint instead. now it's clear that a separate sub-project would be better [13:40] so it could have a mailing list, and commit mail sent there etc [13:40] fluffypony: what are those rules? [13:40] http://cipherdyne.org/psad/docs/config.html [13:41] http://www.cipherdyne.org/psad/docs/fwconfig.html [13:41] yes [13:41] I also tried this: http://bodhizazen.net/Tutorials/psad [13:41] jacobw, cool man now i got boot screen [13:41] thank you ! [13:41] where he adds it to the UFW before.rules - -A ufw-before-input -j LOG --log-level warn [13:41] but that just broke UFW [13:41] and it wouldn't even start [13:42] following bodhizazen's guide? [13:42] tjaalton: what was their argument for collab-maint? not enough packages (or contributors) to warrant a separate group? [13:42] jacobw: yes - that UFW rule is invalid (it seems) [13:44] kklimonda: it would've been ok to create the project, but it was kinda put on hold. no specific arguments [13:46] i just couldn't decide, so went for collab-maint [13:46] right, it works fine anyway untill there are more people who are interested in helping out [13:46] jacobw I found no .exe file for windows [13:47] fluffypony: it seems that problem is with lucid switching to rsyslog [13:48] jacobw: yep - in 8.04 I could just echo -e 'kern.info\t|/var/lib/psad/psadfifo' >> /etc/syslog.conf and then restart sysklogd [13:50] tjaalton: hmm, do you use some different setup for git-buildpackage, or don't use it at all? [13:50] don't use it [13:50] fluffypony: have you added any rules to /etc/rsyslog.d/* ? [13:51] no - 20-ufw.conf 50-default.conf postfix.conf [13:53] greenhost: http://sourceforge.net/projects/wubdepends/files/wubdepends/Initial/wubdepends3.exe/download [13:53] hi [13:55] zaur-ibr: bye [13:55] :-P [13:57] fluffypony: add a rule to rsyslog.d/psad.conf to pipe all messages containing 'firewall' to /var/lib/psad/psadfifo, restart rsyslog and test [13:58] ok doing now [13:59] fluffypony: check /etc/ufw/*.rules to make sure that messages are being logged after all other rules [14:00] fluffypony: and restart ufw, or start it if its not running [14:01] ufw's off atm [14:01] so it's just the iptables rules I've added [14:01] kklimonda: if you have further questions you can email me, or the lists on launchpad [14:03] jacobw: doesn't seem to be working [14:03] I think the problem is with iptables not logging anything [14:04] jacobw how to isntall python terminal for windows? [14:04] I mean, I've got this in my Chain INPUT and Chain FORWARD: LOG all -- anywhere anywhere LOG level warning prefix `firewall ' [14:06] fluffypony: have you restarted rsyslog? [14:06] fluffypony: if not, check /var/log/syslog for 'firewall' messages first [14:06] tjaalton: sure, I will :) [14:07] confirm iptables is logging, then confirm rsyslog is piping correctly to /var/lib/psad/psadfifo [14:08] after that, confirm that psad is reading /var/lib/psad/psadfifo [14:08] * jacobw → back soon [14:08] Feb 12 16:08:02 primary rsyslogd: [origin software="rsyslogd" swVersion="4.2.0" x-pid="7926" x-info="http://www.rsyslog.com"] (re)start [14:10] ja no - syslog has no entries in it [14:10] I mean, it has entries, but no firewall entries :) [14:14] hmmmmm [14:14] I have a feeling that even though UFW is disabled it's rules are still in iptables, and mine are added after the UFW rules [14:14] time to kill of UFW [14:19] hrm [14:19] server crashed [14:19] sigh. [14:30] inaccessible by ssh? [14:34] well that was fun - router decided to restart as well :-P [14:37] ok where was I [14:37] so I removed UFW, and inserted those rules [14:37] but I'm not getting anything in syslog [14:40] fluffypony: which rules? [14:42] the iptables rules [14:43] these ones? http://bodhizazen.net/Tutorials/psad#Configure_psad [14:44] no [14:44] those are borked [14:44] http://glycogen.net/2011/05/05/setup-psad-in-ubuntu/ [14:44] those ones [14:45] ok [14:46] ok - time to reinstall [14:46] then I'll re-evaluate if it still doesn't work [14:46] thanks jacobw [14:46] hang on [14:46] ok [14:47] pastebin iptables -L [14:47] that's the reason I'm reinstalling - I installed Zentyal to play with [14:47] and that has messed up my entire installation [14:48] and apt-get remove zentyal && apt-get autoremove didn't fix it [14:48] clearly they want you to install and never remove it [14:49] it depends how its packaged [14:49] yeah [14:50] aaaah [14:50] did you add the psad.conf rule to rsyslog.d/ ? [14:50] I see it installs a bunch of modules as well [14:50] how nice [14:50] *removes* [14:50] jacobw: yes [14:51] but I'm not seeing stuff get logged to the syslog [14:51] check /var/lib/psad/psadfifo [14:51] so that rule will never kick in [14:52] emkpty [14:52] empty [14:53] ok, if you added the rules from glycogen.net to the top of 50-default.conf it'd explain no messagesg in syslog despite iptables logging [14:54] so, there's a package called pastebinit that provides a script that you can pipe things to, it uploads its input to pastebin and gives back a url [14:54] ok cool [14:55] so the rules shouldn't be at the top of 50-default?? [14:55] *confused* [14:56] if you want messages containing 'firewall' to go to /var/lib/psad/psadfifo and nowhere else, then yes [14:57] there's two rules, the first matches and pipes to psadfifo, the second matches and drops [14:57] aaaaaaah [14:57] that's why I'm not seeing it [14:58] ok commented out the second one [15:00] iptables -L | pastebinit [15:00] hrrrrrrm [15:00] I can't install pastebinit [15:01] no DNS resolution [15:02] brb - [15:04] ok flushed iptables [15:04] http://pastebin.com/5ewUTPRs [15:07] ok, rm psad.conf or remove the lines from 50-default and restart rsyslog [15:11] k [15:14] done [15:15] btw the way I'm testing is nmap -sV -F from a different box [15:16] that should trigger those rules, right? [15:16] everything will be logged [15:16] psad analyses the log [15:17] ok [15:17] tail -f /var/log/syslog [15:17] already checked - nothing [15:17] just postfix [15:17] and some other junk [15:17] I should be seeing tons of stuff being logged [15:18] yes [15:19] ok so [15:19] iptables isn't logging [15:19] and iptables is clearly on - those Zentyal rules locked everything down (in a bad way) [15:20] so maybe iptables is logging to a non-standard locationm? [15:22] boom [15:23] dmesg [15:23] it logs to the kern.* facility === tsimpson_ is now known as tsimpson [15:24] so dmesg is showing the firewall messages [15:24] and rsyslog is dropping them [15:24] see if the message are in /var/log/kern.log [15:25] nope [15:32] hmm [15:33] yeah I know [15:34] i assumed that iptables would log to syslog, which is usually configured to be the case, but rsyslogd might not do this [15:34] yeah - asking in #rsyslog, but it seems like the quietest channel on earth [15:34] rsyslogd uses a module to read from the kernel log and write to syslog [15:34] imklog [15:35] ok [15:35] guys, can anyone help with trim @ ssd raid 1? [15:35] so, `grep -Rl /etc/rsyslod*` [15:35] I added the "discard" option, rebooted the server [15:35] but according to the test I am doing - trim is not working [15:36] rsyslog.conf:$ModLoad imklog [15:37] so it's loaded in rsyslog.conf [15:37] yeah [15:40] great support ... [15:40] Onepamopa: what test are you doing? [15:40] seq 1 1000 > testfile [15:40] hdparm --fibmap testfile [15:40] sync [15:40] jacobw, do you have any idea on postfix/dovecot ? i have problem with pop3 server [15:40] hdparm --read-sector /dev/sda [15:40] then deleting testfile and reading same sector [15:40] samba35: what problem do you have? [15:41] jacobw it should output only zeros after I deleted testfile and sync [15:41] but it outputs the same as first hdparm --read-sector /dev/sda [15:41] i am behind a utm (using proxy /relay for mail) i am using smarthost i am able to send mails but i am not able to get mail [15:42] ok, so added 'discard' to fstab? [15:42] # / was on /dev/md0 during installation [15:42] UUID=327d76c1-f61a-4da1-bd12-502049142982 / ext4 discard,errors=remount-ro 0 1 [15:42] # /var/lib/mysql was on /dev/md1 during installation [15:42] UUID=4f152a4c-7e84-4835-b1b3-5ecfd3f4ce0a /var/lib/mysql ext4 discard,defaults 0 2 [15:43] samba35: how are you testing receive mail? [15:43] samba35: with imap/pop? [15:43] pop [15:43] now idea how to setup a imap [15:43] samba35: postfix logs to /var/log/mail.log [15:44] samba35: dovecot does pop and imap [15:44] ok let me try again from client now i am on server [15:44] i will try from client [15:44] samba35: its likely that postfix isn't delivering properly to dovecot [15:44] jacobw any thoughts? [15:45] when i send mail from gmail to my account mail doesnt get bounce so its comming but i fail to understand where its comming [15:46] fluffypony: test where kern.* messages go with logger [15:48] kern.* -/var/log/kern.log [15:49] according to 50-default in rsyslog.d [15:49] and that seems to match with reality [15:50] Onepamopa: what kernel are you running? [15:51] samba35: check /var/log/mail.log for messages from postfix about delivery [15:56] ok so check [15:56] rock and hard place [15:56] http://toastresearch.com/2011/04/09/packet-logging-with-iptables/ [15:56] "i wasted 2 hours of my life getting this working on a fresh install of unbuntu 10.10. turns out that the default version of rsyslog that you get when you ‘apt-get install rsyslog’ is version 4.x, which has a bug that prevents the logging from being directed correctly to /var/log/iptables.log" [15:56] so I'm going to just upgrade rsyslog, right? [15:56] except [15:56] WARNING: there is a bug with kernels < 2.6.34 causing rsyslog to eat cpu time [15:56] *bangs head against wall* [15:57] i also dislike rsyslog [15:57] I suppose I can replace it with syslog-ng [15:59] are you running 10.10? [15:59] no [15:59] 10.04 LTS [16:00] it's a server after all :-P [16:00] check 50-defaults for iptables > /var/log/iptables.log [16:00] I added to 50-defaults for that [16:01] and then was trying to find out why it wasn't working [16:01] sometimes ok [16:01] er, -sometimes [16:01] check /var/log/messages [16:02] just find out where the iptables messages are going, find the rule in rsyslog.d/ and redirect to psadfifo :) [16:03] if I grep for iptables in /var/log [16:03] all that comes up are entries in dpkg.log [16:03] lol [16:09] holy crap [16:09] upgraded rsyslog [16:09] my kern.log is now 447mb [16:09] 592mb [16:09] the messages might not be tagged with 'iptables' [16:09] rofl [16:11] rsyslog will be writing cached messages now [16:11] yeah no [16:11] it's the bug [16:12] see if it slows, if not add a log chain to iptables and only log certain traffic types etc [16:20] ok [16:21] so I killed rsyslog [16:21] installed syslog-ng [16:21] much better [16:32] ok sweet [16:32] works [16:32] syslog-ng supports named pipes [18:37] Anyone know if the 11.10 64 bit server installer iso has xfs support natively? I was surprised that the 10.04 one does not... [18:52] anyone help? I need to recursively search all folders under a directory, and move all files with a specific extension to another location… Ie, all files under movies with the extension AVI, move to here… where here is my move location… I am thinking mv -R *.avi /media/Goflex/AVI [18:54] find . -name '*.avi' -exec 'mv' '{}' '/media/Goflex/AVI/' ';' [18:58] Thank you [18:59] ok, now how can I rename all the files to remove the first 5 charters but keep everything the same [19:00] rename [19:00] so "01 - HomeVid0001" to "HomeVid0001" [19:00] I have some thousand files… [19:00] if not more [19:00] since when did rename work on one file at a time? [19:01] well that is why I am asking [19:01] and I told you to use rename [19:03] ok…. what is the pearlexp? IDK anything about pearl…. [19:04] regexp [19:05] rename 's/^[0-9]+ - (.*)$/$1/' *.avi [19:05] might do what you want [19:05] ok, I appreciate your help, can you explain that syntax? [19:06] http://lmgtfy.com/?q=regexp [19:08] yes that worked… thank you === aarcane_ is now known as aarcane [20:22] tjaalton: any idea where is ndr_nbt library from? [20:22] (it's one of freeipa dependencies, it's not in our samba4 sources, like other "ndr" libraries) [20:24] ah, apparently it's a new library, not present in alpha17 release - holy molly, it feels like freeipa folks are doing everything to make it harder to build this stuff on systems other than fedora/rhel.. [20:38] what is the correct way to fix this bug/error > PHP Notice: session_start(): ps_files_cleanup_dir: [20:39] a notice is just that, a notice [20:39] it doesn't mean anything is wrong [20:41] its blocking lots [20:41] hmm [20:41] i disagree :P [20:41] i dont like them [20:41] i know i can supress, but it does mean something is not quite right [20:41] the notice itself is not an error, it may be hiding the actual error [20:42] well i read that PHP doesnt have write access to its garbage location? [20:42] on ubuntu [20:42] but the default in ubuntu cleans this folder anyway [20:42] so in this case i guess i can suppress the warning [20:43] error levels go something like this: info, notice, warn, error [20:43] info and notice are just letting you know that things are happening [20:43] warn could be a problem, but could just be a warning, and error means something fatal happened [20:44] if you're getting unexpected behavior, info and notice can lead you to the problem, but they're not direct errors of problems, and won't necessarily lead you to what's wrong by themselves [20:45] it seems messy though, i had a script and was getting tons of 'index not defined PHP errors' i have wrapped some of them in '''s and it fixed the warning [20:45] is that just poor coding>? [20:46] the script is probably written for an older (or possibly newer) version of PHP [20:46] yea i thought that [20:46] so this [20:46] in ubuntu [20:46] the major version points to be careful of are PHP4 to PHP5, 5.0 to 5.2, and 5.2 to 5.3 [20:46] PHP Notice: session_start(): ps_files_cleanup_dir: opendir(/var/lib/php5) failed: Permission denied (13) [20:46] there is no clean way to fix this error ? [20:46] or 'notice' [20:47] I'd think that'd be more than a notice, but you'd have to find more information on that particular issue [20:47] unless it's causing a problem, I would leave it [20:47] but if it is, you could change the permissions on that directory [20:47] be aware though that this could have major security implications [20:47] i would like to fix so i came here :) [20:47] changing permissions also seems a little 'naughty' [20:48] like i said i think there is a cronjob for it in ubuntu ? [20:48] well, that's how you correct 'permission denied' [20:48] but it may be denied for a good reason [20:48] but if ubuntu is already doing the job, maybe i should remove it from the php ? [20:48] i want the 'cleanest' way :) [20:49] there are only two ways [20:49] change permission, or disable the function that's being denied [20:49] this is what causes the error [20:49] session_start(); [20:50] but its PHP usage of this that causes the notice? [20:51] quick google reveals bug 619855 [20:51] Launchpad bug 619855 in php5 "session.gc_probablity=1 in /etc/php5/apache2/php.ini conflicts with permissions on /var/lib/php5" [Low,Won't fix] https://launchpad.net/bugs/619855 [20:53] and from what I'm reading there, it's a problem between upstream and ubuntu's way, and is safe to ignore completely [20:53] if you just want the messages gone, use @session_start(); instead [20:53] should i set gc_prob to 0 ? [20:53] id rather not have it filling up my logs [20:54] they will still occur with that set to 0, just less often, and that actually reduces functionality [20:55] ic [20:56] i like this guys reply > Agustín Fernández (yo-agustinfernandez) wrote on 2010-09-09: [20:56] surely one of the other should be fixed, settings or permissions, via ubuntu/php [20:56] the final reply sums it up best [20:56] it's unfortunate but it provides the best default [20:57] lol [20:57] should i chown the DIR to root:www-data? [20:58] it's a matter of which is worse, breaking security, breaking compatibility, or log noise [20:58] right now there's log noise [20:58] changing permissions break security [20:58] and changing the setting breaks compatibility [20:58] can i 'loosen' them like he suggests [20:58] without breaking security [20:58] not likely [20:58] I would leave it as is, personally [20:59] makes reading logs painful :) [20:59] second choice modifying the script with the @session_start(); [20:59] third choice disabling the feature [20:59] guess i can grep the errors out though [20:59] what does the @ do ? [20:59] suppresses errors for that line [20:59] nice. [20:59] :) [20:59] ty [21:22] hallyn: just noticed a few more services failing because of the lack of net-device-up of 'lo'. I'll upload a new ifupdown including a new upstart job emitting the needed net-device-added and making everything work for containers. [22:06] If i am using a SSL virtualhost, like the default one in ubuntu, so i replace _default_ with serverIP or domain name? [22:08] Tixos: No, why? [22:08] Will you be using multiple vhosts ? [22:14] what should it be? [22:14] the domain name? [22:25] getting some 408 errors, wondering if its related at all to bad config [22:26] but like this, what does it mean [22:26] [12/Feb/2012:16:26:38 -0600] "-" 408 145 "-" "-" [22:32] any ideas? getting quite a few, should i jack up max clients or something? [22:34] keepalive timeout ? [22:36] tum tee tum [22:36] gah, wrong window, sorry === duckydan_ is now known as duckydan [23:27] hi all === n000 is now known as koolhead11 === Jasonn_ is now known as Jasonn [23:47] hello there [23:51] hi fraterm