| [00:26] <frojnd> jmarsden: ok thanx | |
| [00:26] <frojnd> ..whre can I change default port 80 for http ? | |
| [00:27] <frojnd> I need to change it to something above 1000 since my ISP blocks everything under 1000 | |
| [00:28] <frojnd> I found it in apache2/ports.conf :) | |
| [00:28] <jmarsden> frojnd: /etc/apache2/ports.conf | |
| [00:28] <jmarsden> Yup. | |
| [00:30] <jmarsden> You may also need to chaneg it in various VirtualHost lines, see /etc/apache2/sites-available/default for example | |
| [00:31] <frojnd> jmarsden: hm hm :) | |
| [00:43] <owh> When I run aptitude install libvirt-bin, it installs three packages: dnsmasq-base libvirt-bin netcat-openbsd -- but when I run aptitude purge libvirt-bin, it only wants to remove dnsmasq-base libvirt-bin -- this makes no sense. Why is this happening? | |
| [00:46] <owh> Is this a bug in libvirt-bin? | |
| [00:47] <jmarsden> Or a bug in aptitude?? Hard to say. Have you checked all the Depends: and Recommends: stuff in all 3 packages for clues? | |
| [01:00] <owh> jmarsden, I don't see any difference between the depends and rdepends of dnsmasq-base and netcat-openbsd. | |
| [01:01] <jmarsden> Strange... OK, let me see if I can duplicate the bug here... | |
| [01:14] <owh> jmarsden, I've compared /var/lib/aptitude/pkgstates in the three states, that is, before the install, during the time that libvirt-bin is installed and after it has been purged. netcat-openbsd looks the same as dnsmasq-base, except in the final state -- because it's still installed. Very curious. | |
| [01:15] <jmarsden> Indeed. I wonder if it could have anything to do with the way the netcat packages use alternatives somehow...? But I don't really see why that should affect this. | |
| [01:17] <frojnd> oh nose :s | |
| [01:17] <frojnd> how can I update CA certificate ? | |
| [01:20] <frojnd> anyone ? | |
| [01:22] <owh> jmarsden, were you able to reproduce this? | |
| [01:22] <jmarsden> owh: Yes, I reproduced the issue here (Intrepid on amd64) | |
| [01:23] <ball> my head hurts. so. much. | |
| [01:23] <jmarsden> frojnd: Update a CA cert? Normally you would just create a new one. | |
| [01:24] * owh hands ball an ice-cream for the pain. | |
| [01:27] <owh> jmarsden, I'm not sure where to report the issue first. I'd rather not report a bug to everything. What do you think of an email to ubuntu-devel-discuss as a start? | |
| [01:28] <jmarsden> You could do that; I'd be more tempted to just report it as an aptitude bug, indicating you are not sure if it is really a bug in aptitude or one of the packages concerned, and let whoever triages it decide whether that is correct or not. | |
| [01:30] <owh> jmarsden, cool, I can do that instead. | |
| [01:34] <owh> jmarsden, FYI the bug exists - I'm adding a comment: Bug #305504 | |
| [01:34] <uvirtbot`> Launchpad bug 305504 in aptitude "aptitude does not automatically remove unused packages" [Undecided,Confirmed] https://launchpad.net/bugs/305504 | |
| [01:37] <frojnd> ...does anyone know how can I update cert list ? | |
| [01:38] <jmarsden> frojnd: Read https://help.ubuntu.com/community/OpenSSL for general info on certificates and Apache? | |
| [01:39] <frojnd> just CA | |
| [01:40] <jmarsden> frojnd: That page describes creating your own local CA, and using it to create certs, etc... | |
| [01:43] <frojnd> jmarsden: very nice :) | |
| [01:43] <frojnd> it solved my problem :) | |
| [01:43] <jmarsden> Good! | |
| [02:37] <fowlduck> oh my kittens, 137 ppl | |
| [02:40] <techsupport> how can i uninstall perl without the other packages ? | |
| [02:40] <fowlduck> i'm not sure how to google this, so i suppose i'll describe my situation and maybe someone can give me the terminology for what i'm looking for | |
| [02:41] <Kamping_Kaiser> techsupport, you dont. | |
| [02:41] <techsupport> cause when i type sudo apt-get remove perl it wants to uninstall apache2, mysql, total is 182 megs | |
| [02:42] <Kamping_Kaiser> you dont uninstall perl. its like uninstalling oxygen. | |
| [02:42] <techsupport> i need to downgrade | |
| [02:42] <fowlduck> I'm running a rails application that's using postfix for delivery of mail, specifically just sendmail. what is it called when your mail server only does that? local delivery only? i'd like to figure out how to disable any other features to keep it locked down | |
| [02:42] <Kamping_Kaiser> techsupport, huh? | |
| [02:43] <techsupport> the version of perl that i have is 5.10 and i need 5.8 | |
| [02:44] <Kamping_Kaiser> techsupport, whysthat? | |
| [02:44] <Kamping_Kaiser> fowlduck, the exim4 installer calls it "local delivery". not sure about posfix though. | |
| [02:44] <techsupport> because its not compatible with utf8 | |
| [02:44] <fowlduck> Kamping_Kaiser: hm, so i should reconfigure postfix for local delivery only, maybe | |
| [02:45] <Kamping_Kaiser> fowlduck, if its not already set to do that, yeah. | |
| [02:45] <Kamping_Kaiser> techsupport, perl 5.10 isnt? nonsense | |
| [02:45] <fowlduck> Kamping_Kaiser: i think i configured it for "Internet Site" | |
| [02:45] <fowlduck> since it's a site on the internet ;) | |
| [02:46] <Kamping_Kaiser> internet site probably means connects to other hosts via smtp directly. | |
| [02:47] <fowlduck> Internet Site: Mail is send and received directly using SMTP | |
| [02:47] <fowlduck> Local only: The only delivered mail is the mail for local users. There is not network. | |
| [02:47] <Kamping_Kaiser> yay, i got something right ;P | |
| [02:48] <fowlduck> i don't think local only is right, based on that description | |
| [02:48] <fowlduck> hm, maybe i should try exim | |
| [02:48] <Kamping_Kaiser> what do you want the mail server to do? | |
| [02:48] <fowlduck> just allow an application on the system to use sendmail to send email | |
| [02:48] <fowlduck> no receiving, no relay, no remote sendinf | |
| [02:48] <fowlduck> sending* | |
| [02:49] <Kamping_Kaiser> local should do what you want, as long as the mail is delivered locally | |
| [02:49] <fowlduck> i know so little about how email servers work *sigh* | |
| [02:49] * Kamping_Kaiser isn't sure he made sense there | |
| [02:49] <fowlduck> define 'delivered locally' | |
| [02:50] <Kamping_Kaiser> to a user on the system | |
| [02:50] <fowlduck> it's delivered to emails outside of the domain, if that's what you mean by not delivered locally | |
| [02:50] <fowlduck> yeah, it's not just locally | |
| [02:50] <fowlduck> for example, a user signs up on the site, we send out a confirmation email | |
| [02:51] <fowlduck> maybe i don't need postfix or exim, just sendmail? | |
| [02:51] <Kamping_Kaiser> have a look at ssmtp | |
| [02:51] <Kamping_Kaiser> as long as your network connection stays up, its pretty funky | |
| [02:51] <Kamping_Kaiser> (it doesnt hold unsendable messages, it drops them). | |
| [02:52] <fowlduck> yeah, someone already recommended that one, not sure if i want to go with it | |
| [02:53] <techsupport> how can i downgrade my perl version ? | |
| [02:54] <Kamping_Kaiser> with a world of pain - I'm certainly not going to try and help you do it | |
| [02:54] <fowlduck> Kamping_Kaiser: that could be an issue, too, dropping unsendable messages :/ | |
| [02:54] <fowlduck> well, potentially | |
| [02:54] <techsupport> ubuntu server 8.10 , from perl v.5.10 to perl v.5.8 | |
| [02:54] <fowlduck> we usually just use a background process with gafyd using smtp straight from the app | |
| [02:54] <fowlduck> but this is a legacy app :/ | |
| [02:55] <fowlduck> no time to put all that in and track emails that are sent | |
| [02:55] <Kamping_Kaiser> perhaps search the package repos for a simple smtp sender that doesnt drop mail? otherwise setting postfix to listen on loopback and send via $public is your best bet | |
| [02:58] <fowlduck> hmm, cool, found some good exim tutorials | |
| [02:58] <fowlduck> http://newbiedoc.sourceforge.net/networking/exim.html | |
| [03:33] <fowlduck> Kamping_Kaiser: thanks, btw. i just went with postfix and set mynetworks to 127.0.0.1 | |
| [03:33] <fowlduck> supposed to only relay email for that ip then | |
| [03:33] <Kamping_Kaiser> fowlduck, cool. have fun, hope it works :) | |
| [03:33] <fowlduck> same ;) | |
| [06:48] <kirkland> nijaba: thanks! i just merged and tested | |
| [06:49] <kirkland> nijaba: couple of minor tweaks | |
| [06:49] <kirkland> nijaba: renamed screen-install and screen-remove to screen-launcher-install and screen-launcher-uninstall | |
| [06:49] <kirkland> nijaba: seems more accurately named | |
| [06:50] <kirkland> nijaba: i just released 1.6 to my ppa's and to jaunty universe | |
| [06:50] * kirkland going to bed now :-) | |
| === CrummyGummy is now known as everyoneelse | |
| === everyoneelse is now known as CrummyGummy | |
| [08:35] <uvirtbot`> New bug: #316283 in bind9 (main) "bind9-dlz-mysql - Please make a package with this option turned on" [Undecided,New] https://launchpad.net/bugs/316283 | |
| [11:16] <kraut> moin | |
| [11:33] <PC_Nerd> Hi. Attempting to send emails from batch scripts/crons etc via mail (exim4 etc). Dont want to run a mailserver, and have port 25 open.... any ideas why Im not getting any emails from the mail command? (about to post logs). | |
| [11:40] <PC_Nerd> http://pc_nerd.pastebin.com/m601147ca thats one log entry from one mail command ( the command used included) | |
| [11:40] <PC_Nerd> any ideas? | |
| [12:09] <jobrien> I am having issues with a clean install of 8.10 and 8.04 on a Dell 2500 with PERC Controllers. Has anyone had any luck with the PERC based systems? | |
| [12:10] <Kamping_Kaiser> PC_Nerd, "Mailing to remote domains not supported". sounds like your config needs tweaking | |
| [12:10] <_ruben> jobrien: i installed both 8.04 and 8.10 just fine on dell pe1950 and pe2950 | |
| [12:11] <_ruben> perc/5 and perc/6 i think | |
| [12:11] <PC_Nerd> ok - where should I be looking to "tweak" it? | |
| [12:12] <jobrien> _ruben: I saw people were having some luck with those. My OS drive is in RAID 1 and the install goes perfect then I get GRUB 17 at boot. Been screwing with this for 4 days. | |
| [12:12] <_ruben> jobrien: never had any grub problems with em | |
| [12:13] <jobrien> _ruben: Thanks man. I"ll just keep digging. I RHL boots fine but I really would like to stick with a debian base. | |
| [12:39] <Phil_> hello, does anybody know how to mount an hfsx filesystem in linux? | |
| [12:40] <uvirtbot`> New bug: #300265 in libapache2-mod-perl2 (main) "armel build failure (package not yet in the archive)" [High,Fix released] https://launchpad.net/bugs/300265 | |
| [12:41] <Kamping_Kaiser> didnt realise ubuntu had an armel port | |
| === BrunoX1ambert is now known as BrunoXLambert | |
| [13:43] <heath|work> what's the best way in a script to test if a drive is mounted to a specific location? Right now I am putting a file in the mount folder called NOT_MOUNTED then testing for it | |
| [13:44] <sommer> heath|work: probably the mount command, then you could use grep or awk on the output | |
| [13:44] <heath|work> is there a way for mount to show by uuid?\ | |
| [13:46] <sommer> heath|work: mmm, not that I see, but there may be another utility that can | |
| [13:50] <heath|work> I will look into it, thanks | |
| [14:03] <ScottK> FYI, https://edge.launchpad.net/~ubuntu-dev/+polls for any developers. | |
| === henkjan_ is now known as henkjan | |
| === [1]jobrien is now known as jobrien | |
| [14:34] <kirkland> nijaba: around? | |
| [14:40] <jdstrand> kirkland, nijaba: with the latest screen-profiles, help (F9) doesn't appear to work | |
| [14:40] <kirkland> jdstrand: ? | |
| [14:40] <kirkland> jdstrand: what version? | |
| [14:41] <jdstrand> kirkland, nijaba: I also noticed that after the update I did 'select-screen-profile' then launched screen, I got a flash of blue before screen started. I deleted all my (previously unmodified) .screenrc* files and that seemed to go away, but help still doesn't work | |
| [14:43] <jdstrand> kirkland: version is 1.7-0ubuntu1 | |
| [14:45] <kirkland> jdstrand: testing ... | |
| [14:45] <jdstrand> this is in gnome-terminal, if that makes a difference | |
| [14:46] <kirkland> jdstrand: f9 is working for me, in gnome-terminal | |
| [14:46] <kirkland> jdstrand: 1.7-0ubuntu1 | |
| [14:46] <kirkland> jdstrand: try removing ~/.screen-profiles-helper | |
| [14:48] <frojnd> Just curious why a user can't have autocmopletion with tab ? | |
| [14:48] <frojnd> Do i have to se it manually ? | |
| [14:48] <jdstrand> kirkland: I don't have that file | |
| [14:49] <frojnd> when I do du -hs /media[tab] it will throw me a few spaces forward... I'd like tab for autocompletion for a user ... | |
| [14:49] <kirkland> jdstrand: ah, that looks to be the problem | |
| [14:49] <frojnd> How can I do that ? | |
| [14:49] <kirkland> jdstrand: nijaba's toggle code didn't account for that file missing | |
| [14:50] <kirkland> jdstrand: i'll fix | |
| [14:56] <jdstrand> kirkland: cool | |
| [14:59] <frojnd> ...looks like I don't have a bash as a user ... | |
| [14:59] <frojnd> how come only first user which is also root has bash ? | |
| [14:59] <frojnd> how can I give bash to everyone else ? | |
| [15:06] <kirkland> jdstrand: okay, i think i have it | |
| [15:06] <kirkland> jdstrand: would you mind verifying it? | |
| [15:07] <jdstrand> kirkland: ok | |
| [15:07] <kirkland> jdstrand: jaunty, right? | |
| [15:07] <jdstrand> kirkland: where is it? | |
| [15:07] <jdstrand> kirkland: oh yes, jaunty | |
| [15:07] <kirkland> jdstrand: i'm pushing to my ppa now | |
| [15:10] <kirkland> jdstrand: i'll ping you as soon as the dsc is available, and you can build it | |
| [15:10] <kirkland> jdstrand: faster than ppa will build ;-) | |
| [15:10] <jdstrand> kirkland: ok | |
| [15:10] <kirkland> jdstrand: i'll upload to universe as soon as you confirm | |
| [15:10] <kirkland> jdstrand: dget https://edge.launchpad.net/%7Ekirkland/+archive/+files/screen-profiles_1.8-0ubuntu1~ppa1.dsc | |
| [15:14] <jdstrand> kirkland: \o/ | |
| [15:14] <jdstrand> kirkland: working now | |
| [15:14] <jdstrand> kirkland: thanks! | |
| [15:14] <kirkland> jdstrand: cool, it need a couple of try/except cases | |
| [15:14] <kirkland> jdstrand: to handle that config file not existing | |
| [15:14] <frojnd> for everyone else that had similar problem like me... | |
| [15:15] <kirkland> jdstrand: i'll upload now | |
| [15:15] <Shoopuf> Is there some sort of script or program I can use that will aggregate all my apache/whatever logs together so I can analyze them to see if anyone has been suspiciousing? | |
| [15:15] <kirkland> Shoopuf: there are a few, i think | |
| [15:15] <Shoopuf> suspiciousingness* | |
| [15:15] <frojnd> if ur new to shell and u just created a new user and has /bin/sh instead of a /bin/bash shell all u have to do is to change it's shell like this: chsh -s /bin/bash username | |
| [15:16] <Shoopuf> frojnd: chsh -s /bin/bash shoopuf? ok thank you | |
| [15:16] <kirkland> Shoopuf: i use awstats | |
| [15:16] <Shoopuf> kirkland: ok i'll try that, thanks | |
| [15:19] <frojnd> How can I check if my external disk supports ext2 or ext3,... it's 1TB and I'm not fine with just vfat32... | |
| [15:19] <frojnd> Is it even possible ' | |
| [15:28] <heath|work> frojnd, why would it not support ext3? | |
| [15:32] <RainCT> Hi | |
| [15:33] <RainCT> Does someone know how I can enable mod_userdir for those users in a specific group? | |
| [15:36] <Jeeves_> RainCT: http://httpd.apache.org/docs/2.2/mod/mod_userdir.html | |
| [15:36] <Jeeves_> It seems that you cannot do that | |
| [15:36] <RainCT> :( | |
| [15:47] <Jeeves_> Who is using kvm/libvirt with nfs storage? | |
| [16:02] <soren> Jeeves_: A number of people... Why? | |
| [16:03] <Jeeves_> soren: I get issues when migrating | |
| [16:03] <Jeeves_> it looks like the image file is still locked by the source host | |
| [16:03] <Jeeves_> Migration is successfull | |
| [16:03] <Jeeves_> the domain starts pinging after about four seconds (it's a 100mbit link) | |
| [16:04] <soren> Probably a bridging issue. | |
| [16:04] <soren> Try setting fd for your bridge to 0. | |
| [16:04] <Jeeves_> soren: How do you mean? | |
| [16:04] <soren> You're using bridged networking, right? | |
| [16:05] <Jeeves_> Yes | |
| [16:05] <soren> brctl setfd <name of your bridge> 0 | |
| [16:06] <Jeeves_> oki | |
| [16:06] <Jeeves_> lets ry | |
| [16:06] <Jeeves_> +t | |
| [16:13] <Jeeves_> That doesn't seem to help | |
| [16:13] <Jeeves_> But I'll reconfigure it and restart the boxes, just to be sure | |
| [16:16] <Jeeves_> 17:15 < bitrot> SVN commit by marks on repository cfengine, revision 11248: Set bridge_fd < soren> brctl setfd <name of your bridge> 0 | |
| [16:16] <Jeeves_> Tadaaa.wav | |
| === _jmedina is now known as jmedina | |
| [16:22] <_Cid> question ...why is 'sudo -i' considered "better to use" than 'sudo /bin/su' ? | |
| [16:23] <soren> Jeeves_: Any better? | |
| [16:24] <ph8> hi all - any recommendations on the best mass-market mailing list manager to use on ubuntu? (e.g. for box users to make their own mailing lists with etc) | |
| [16:25] <soren> mailman? | |
| [16:26] <ph8> cool | |
| [16:26] <ph8> that's what i was thinking of | |
| [16:26] <ph8> thought i'd ask first | |
| [16:27] <Jeeves_> soren: Nope | |
| [16:27] <Jeeves_> bit-beheer@ubuntu:~$ sudo -i | |
| [16:27] <Jeeves_> -bash: /usr/bin/sudo: Input/output error | |
| [16:28] <Jeeves_> soren: http://pastebin.ubuntu.com/103979/ | |
| [16:31] <jmedina> ph8: I like phplist | |
| [16:32] <jmedina> it has better features for only sending mailing, open/click count | |
| [16:32] <jmedina> support for text-plain and html templates, and scheduling | |
| [16:33] <jmedina> clean interface for opt-in/opt-out | |
| [16:33] <ph8> i'll take a look, thanks | |
| [16:34] <soren> Jeeves_: What's that sudo thing? | |
| [16:36] <soren> Jeeves_: I'm curious, though... You got kvm migration running under libvirt how? | |
| [17:05] <frojnd> How come when I execute screen a user doesn't have a bash but sh ? if I do echo $SHELL when not in screen I get a nice /bin/bash but when in a screen and executing echo $SHELL I get /bin/sh ? | |
| [17:06] <frojnd> How can I fix this so when in screen a user will also have bash ? | |
| [17:21] <heath|work> if a bash script, can you do an or something like if [ -e $FILE || -d $FILE ] ? | |
| [17:27] <nixphreak> should I need a broadcast address if I am only using one server using a static ip from an ISP ? | |
| [17:39] <jmarsden|work> heath|work: man test describes all the operators you can use. I think you want [ -e "$FILE" -o -d "$FILE" ] | |
| [17:54] <heath|work> jmarsden|work, thanks... I found that if [[ -e $FILE || -d $FILE ]] works too | |
| [17:56] <jmarsden|work> No problem. I'm showing my age, [[ is bash-specific and I think didn't exist when I learned shell programming -- but yes, that will work fine too. | |
| [18:23] <Chipzz> heath|work: [ actually is a command with an exit status (there's also a binary called [) | |
| [18:23] <Chipzz> hence what you should be doing is [ sometest ] || [ someothertest ] | |
| [18:24] <Chipzz> but do not use [[ ; that's not portable at all, and there is no reason at all to prefer it over [ | |
| [18:26] <Chipzz> || is actually a way of constructing a bash pipeline (please refer to "man bash" for more info), and seperates 2 seperate commands | |
| [18:45] <kees> jdstrand: I like the ec2sec script. I was pondering renaming "start" to "launch", and making "start" call "launch" and then wait for the new instance to leave "pending". (i.e. once "start" finishes, the instance is _actually_ running.) | |
| [18:46] <jdstrand> kees: that sounds like a nice improvement | |
| [18:46] <jdstrand> kees: there are surely many more :) | |
| [18:50] * zul is working on new ec2 crack | |
| [18:51] <mathiaz> jdstrand: kees: so your main usage of ec2 is for testing? | |
| [18:52] <kees> mathiaz: yes, though it's still rather informal. | |
| [18:52] <kees> mathiaz: mostly just trying things out, seeing how it operates, etc | |
| [18:53] <mathiaz> kees: right - the main issue I'm facing now is the step once you've launched the image: customize your test environemnt | |
| [18:54] <jdstrand> mathiaz: yeah | |
| [18:54] <mathiaz> kees: on my server I have {hardy,intrepid,...}-base images that I clone (via lvm snapshots) when creating new vm for testin | |
| [18:54] <jdstrand> kees: though, with the stuff you talked about with cr3, some of our environment setup and testing might become easier | |
| [18:55] <jdstrand> mathiaz, kees: something that is cool about ec2 that does fit in nicely is that on launch we have a pristine image | |
| [18:56] * kees nods | |
| [19:37] <ball> mathiaz: Can't make tomorrow's meeting, I'll be at work sadly. | |
| [19:40] <ball> ...not that I had anything useful to contribute anyway mind ;-) | |
| [19:42] <stickystyle> Does anyone know of a way to push a cron job out to all of my users, either creating a new crontab for them (if one doesn't exist) or appending to their existing crontab? | |
| [19:47] <jmedina> probably appendint the job to /var/spool/cron/crontabs/usrname | |
| [19:48] <jmarsden|work> But change the time it runs at for each user... if not you will try to run all of those cron jobs at the same time, which is probably not good for performance ;) | |
| [19:51] <jmedina> well you need to do some shell scriptting to automatically update the time for each user | |
| [20:01] <stickystyle> jmedina: yeah, thats what I thought at first also, but right at the top of each crontab is "DO NOT EDIT THIS FILE - edit the master and reinstall." | |
| [20:02] * stickystyle shrug | |
| [20:03] <jmarsden|work> stickystyle: You can use crontab -e to edit those files if you want, maybe set $EDITOR to sed or whatever and play games that way... | |
| [20:03] <jmarsden|work> But as long as you know noone else is editing the crontabs when you are, appending to the files should be OK. | |
| [20:04] <stickystyle> Hum. okay, thanks for the info. | |
| [20:05] <stickystyle> perhaps that warning is there just for the situation you described of the user perhaps editing the file at the same time...? | |
| [20:06] <jmarsden|work> I'm not 100% sure. There may be some other reason... a test on one user would be smart before editing all of them... | |
| [20:06] <stickystyle> for sure. | |
| [20:08] <jmarsden|work> Looks like crontab -e also does sanity checks on the format of the lines in the file... | |
| [20:20] <erichammond> zul: What is the "new EC2 crack" you're working on? | |
| [20:21] <zul> just some fixes and updated kernel etc | |
| [21:08] <erichammond> zul: Yay. | |
| [21:08] <erichammond> I'm sending out my opinions on prioritization of the ubuntu-on-ec2 bugs. | |
| [21:09] <erichammond> Do you assign tickets when you are working on them? | |
| [21:19] <rickross> anyone updated to the 2.6.28 kernel? | |
| [21:20] <zul> erichammond: i should be doing alot of them this week hopefully | |
| [21:38] <rickross> I have a dir of aprox. 56 GB on an -drive RAID5 with a lot of available spae (linux software raid) | |
| [21:38] <Jeeves_> soren: You awake? | |
| [21:39] <rickross> are there any tricks I can use to get it to copy this dir to a backup copy on the same RAID? | |
| [21:39] <rickross> I have lots of cores and lots of ram available, but I don't think "cp -a" takes advantage of that at all | |
| [21:40] <Jeeves_> rickross: cp usually waits for your disks | |
| [21:40] <Jeeves_> not cpu/ram | |
| [21:40] <uvirtbot`> New bug: #316534 in bacula (universe) "instalation faild, but synaptick still admids the combabilety" [Undecided,New] https://launchpad.net/bugs/316534 | |
| [21:41] <rickross> the copy speed we're seeing is only about 70 MB/s, although the RAID will read a sustained 700 MB/s and write a sustained 300 MB/s | |
| [21:42] <rickross> I think the issue is that it is reading and writing on the same RAID, so it's changing gears all the time to copy files | |
| [21:43] <Jeeves_> rickross: that will decrease your performance indeed | |
| [21:43] <Jeeves_> And a 300MB/sec write? | |
| [21:43] <Jeeves_> What kind of raidset is that? | |
| [21:44] <rickross> 8 1Tb WD Caviar drives in a RAID5 | |
| [21:44] <rickross> we're using an LSI dumb controller that gives us 8 additional SATA ports | |
| [21:45] <Jeeves_> And how long can you write 300MB/sec? | |
| [21:45] <rickross> a long time - we tested for 100 Gb | |
| [21:46] <rickross> but I think it's an entirely different matter to do a recursive copy of a big directory with lots of subdirs | |
| [21:46] <rickross> it actually peaked out at about 400 MB/s for writing, but that's for big blocks and big streams of continuous writing | |
| [21:47] <Jeeves_> Yeah | |
| [21:47] <rickross> so real-world recursive dir copy on the same RAID is showing a big difference :( | |
| [21:47] <Jeeves_> So you'll do about 100MB in real world | |
| [21:48] <rickross> we have a lot of ram available, too, so I wondered if there was some snazzy copy tool that could load a ton of stuff into ram, then pump it out to disk again | |
| [21:48] <rickross> instead of churning back and forth for read/write | |
| [21:48] <Jeeves_> rickross: try 'cp' :) | |
| [21:49] <rickross> jeevs - thx :) | |
| [21:50] <rickross> I didn't see the "cp" option that says "use 16 Gb ram for an input buffer" to avoid switching back and forth so much | |
| [21:50] <Jeeves_> It wouldn't matter for your speed | |
| [21:50] <Jeeves_> It still won't complete untill the last sync() on your disks :) | |
| [21:51] <jmarsden|work> rickross: You could set up a 56GB ram disk and cp to that, then from there to your ultimate destination... if you have that much RAM ;) | |
| [21:53] <jmarsden|work> I suspect the real issue is that your benchmarks for disk speed do not represent the kind of workload you really have and flattered the array... | |
| [21:55] <rickross> jmarsden - undoubtedly | |
| [21:56] <rickross> although there may exist tools which are designed to accelerate this kind of same-device backup | |
| [21:56] <rickross> I just wondered if anyone knew of one | |
| [21:56] <rickross> nslookup mail.hpyle.net | |
| [21:56] <rickross> oops | |
| [21:57] <Deeps> maybe a tool that was originally designed for copying data to multiple floppy disks | |
| [21:57] <Deeps> only instead of floppy disks, it's ram disks | |
| [22:00] <mathiaz> kirkland: I'm working on the Triagger section of the GettingInvolved page for the ServerTeam | |
| [22:00] <kirkland> mathiaz: good for you! | |
| [22:00] <mathiaz> kirkland: and trying to come up with easy steps to follow for new comers. | |
| [22:00] <mathiaz> kirkland: what do you think about advising potential contributors to look at bug in New,Undecided first? | |
| [22:01] <kirkland> mathiaz: definitely... i think moving a bug from "New" to "Confirmed" is a really important first step | |
| [22:01] <kirkland> mathiaz: or marking "Incomplete" if more info is required | |
| [22:01] <kirkland> mathiaz: i tend to focus on Confirmed/Triaged bugs first, as a developer | |
| [22:02] <mathiaz> kirkland: right - should confirmed or triagged have a reproducible test case? | |
| [22:02] <mathiaz> kirkland: the bug squad has the following wiki page: https://wiki.ubuntu.com/Bugs/HowToTriage | |
| [22:03] <mathiaz> kirkland: but it seems a bit long for a first time/potential contributor | |
| [22:03] <kirkland> mathiaz: yes, of course. to move beyond "new", at least someone else should be able to reproduce the issue, and document who to do it (if it's not already in the report) | |
| [22:04] <jmarsden|work> mathiaz: Are there really enough differences between bugs in general and Server bugs that there is a need for a different "HowToTriage" page just for the ServerTeam? | |
| [22:05] <mathiaz> jmarsden|work: I don't think so | |
| [22:05] <kirkland> mathiaz: i agree | |
| [22:05] <kirkland> mathiaz: but i think something like a blogpost would be good | |
| [22:05] <kirkland> mathiaz: calling for more triagers to look at server bugs | |
| [22:05] <kirkland> mathiaz: and giving simple instructions on what that means | |
| [22:05] <kirkland> mathiaz: pointing to the wiki page for more detailed info | |
| [22:06] <mathiaz> jmarsden|work: I'm just updating the section in the GettingInvolved page to make sure potential contributor can get started as easily as possible | |
| [22:06] <jmarsden|work> So... the ServerTeam Traige section could just point ServerTeam contributors to the existing one? | |
| [22:06] <mathiaz> jmarsden|work: right | |
| [22:06] <mathiaz> jmarsden|work: that's already the case from the KnwoledgeBase wiki page | |
| [22:07] <mathiaz> jmarsden|work: where a lot of ressources from the Bugsquad are referenced | |
| [22:07] <mathiaz> I'm mainly thinking about getting potential contributors started as easily as possible - and then give them ressources to read up on. | |
| [22:08] <jmarsden|work> OK, from your questions I thought you were writing a server-specific equivalent of HowToTriage | |
| [22:15] <mathiaz> bdmurray: hi - question about the BugSquad wiki page | |
| [22:15] <mathiaz> bdmurray: I'd like to include the section "Triaging Bugs" from https://wiki.ubuntu.com/BugSquad/GettingInvolved | |
| [22:16] <mathiaz> bdmurray: into the ServerTeam GettingInvolved page | |
| [22:16] <mathiaz> bdmurray: I'd rather not just copy and paste the text as I'd miss all the updates done to the section | |
| [22:16] <mathiaz> bdmurray: the text is already duplicated on https://wiki.ubuntu.com/HelpingWithBugs | |
| [22:17] <mathiaz> bdmurray: do you know of a simple way to do that with moinmoin? | |
| [22:21] <bdmurray> mathiaz: Hi! Yes I do. You can look at Bugs/Tags for an example or ... <<Include(DebuggingOpenOffice, ,5,from="= Bug Tags =", to="= Debugging procedure =")>> | |
| [22:22] <bdmurray> mathiaz: also see http://moinmoin.wikiwikiweb.de/HelpOnMacros/Include | |
| [22:24] <mathiaz> bdmurray: thks :) | |
| [22:28] <mathiaz> bdmurray: I've seens some triaggers moving a bug to confirmed stating that there has been enough comments in the bug as proof of confirmation. Shouldn't a triagger set the status to confirmed if he/she is able to reproduce the bug locally? | |
| [22:30] <bdmurray> Not everyone knows how to change a bug's status or when a bug should be Confirmed. It isn't discoverable outside of Launchpad, subsequently I think having triagers confirming bugs for people who don't know how is fine. | |
| [22:36] <mathiaz> bdmurray: only members of the bugsquad team can set bug statuses right? | |
| [22:37] <bdmurray> mathiaz: no, that's not correct. any one can set any bug status except for Triaged or Won't Fix. | |
| [22:37] <bdmurray> mathiaz: Triaged and Won't Fix are restricted to members of Ubuntu Bug Control | |
| [22:38] <mathiaz> bdmurray: ah - right. | |
| [22:38] <mathiaz> bdmurray: it's the importance that can only be set by members of the bugcontrol team | |
| [22:39] <bdmurray> Yes, that is correct. | |
| [23:23] <mathiaz> kirkland: does screenbin support multiple guest options? | |
| [23:23] <kirkland> mathiaz: yup | |
| [23:23] <mathiaz> kirkland: ie screenbin --guest kirkland --guest mathiaz? | |
| [23:23] <kirkland> mathiaz: yes, precisely | |
| [23:23] <kirkland> mathiaz: says so in the usage() statement | |
| [23:24] <mathiaz> kirkland: ok - I'd mention that in your blog post example. I didn't grasp how the sharing between multiple people work right away. | |
| [23:25] <kirkland> mathiaz: refresh http://blog.dustinkirkland.com/2009/01/screenbin-like-pastebin-but-for-screen.html | |
| [23:26] <mathiaz> kirkland: ah ok. self is automatically added. | |
| [23:26] <kirkland> mathiaz: right | |
| [23:26] <mathiaz> kirkland: that's the part I was missing. | |
| [23:26] <mathiaz> kirkland: since you are kirkland I thought you had to specify your login when using screenbin | |
| [23:27] <kirkland> mathiaz: see: | |
| [23:27] <kirkland> # This could be parameterized for non-standard keypairs | |
| [23:27] <kirkland> for i in id_rsa.pub id_dsa.pub; do | |
| [23:27] <kirkland> if [ -r "$HOME/.ssh/$i" ]; then | |
| [23:27] <kirkland> my_ssh=`cat "$HOME/.ssh/$i"` | |
| [23:28] <kirkland> ... | |
| [23:28] <mathiaz> kirkland: right - I haven't read the code. I was just reading your blog post. It's the first explaination of what screenbin is. And I failed to see how the sharing was working. | |
| [23:30] <kirkland> mathiaz: gotcha, i'll edit one more line in, clarifying | |
| [23:31] <kirkland> mathiaz: okay, update published | |
| [23:39] <techsupport> can someone help me downgrade from perl 5.10 to 5.8 | |
| [23:41] <RainCT> techsupport: Hardy has 5.8 | |
| [23:42] <frojnd> How how how :) | |
| [23:43] <frojnd> I have a little problem. When using screen as normal user I loose bash and I get sh. How come ? | |
| [23:43] <frojnd> I have bash if I'm not using screen... | |
| [23:44] <frojnd> How can I fix this so a user will have bash no matter if he/she uses screen or not ? | |
| [23:46] <techsupport> RainCT, can i transfer my website and database to hardy ? | |
| [23:49] <RainCT> techsupport: yes, that shouldn't be much of a problem | |
| [23:49] <RainCT> techsupport: but first you could try if you can install Hardy's package in Intrepid or whatever you use.. | |
| [23:49] <RainCT> why do you need an older version? perhaps even manually installing perl 5.8 into /usr/local will do? | |
| [23:50] <techsupport> RainCT, if you dont mind reading my post on hlstatsx forum http://www.hlxcommunity.com/forums/viewtopic.php?f=1&t=377 | |
| [23:51] <techsupport> RainCT, if you can help me with this i would really appreciate it | |
| [23:52] <techsupport> using characters not in standard ASCII, it will crash | |
| [23:52] <techsupport> RainCT, please help me :) | |
| [23:53] <RainCT> techsupport: have you read the last post there? | |
| [23:53] <techsupport> RainCT, yes I have | |
| [23:54] <techsupport> RainCT, you mean this ? Basically, whenever someone joins your server with a name that is using characters not in standard ASCII, it will crash. There are many posts here about it, has to do with the MySQL module. | |
| [23:54] <RainCT> techsupport: yes. that guy suggests that it's a problem with MySQL rather than with perl | |
| [23:54] <techsupport> RainCT, no its perl , thats why they dont support perl 5.10 related problems | |
| [23:55] <techsupport> RainCT, the developers dont support 5.10 related questions, they just say use 5.8 perl | |
| [23:55] <RainCT> techsupport: ok. then you could get perl 5.8 from http://www.cpan.org/src/perl-5.8.9.tar.gz, compile and install it into /usr/local | |
| [23:56] <RainCT> then change that script to use /usr/local/bin/perl instead of /usr/bin/perl | |
| [23:58] <RainCT> i'm off now, good night | |
| [23:58] <RainCT> techsupport: and good luck getting that to work :) | |
| [23:58] <techsupport> RainCT, thanx man | |