| MACscr | im running ubuntu 14.04 and spamassasin 3.4. Seems that spamassassin isnt creating the /var/run/spamassassin folder when the service tries to start. If i manually create it and start the service again, everything is fine. Any suggestions? | 00:01 |
|---|---|---|
| === sz0 is now known as sz0` | ||
| === Guest39780 is now known as esde | ||
| === Lcawte is now known as Lcawte|Away | ||
| === thesheff17_ is now known as thesheff17 | ||
| stoned | is it possible to upgrade to 14 from 10? | 02:48 |
| stoned | from 10.04 lts to 12 or 14? | 02:48 |
| cfhowlett | !eolupgrade|stoned | 02:49 |
| ubottu | stoned: End-Of-Life is the time when security updates and support for an Ubuntu release stop, see https://wiki.ubuntu.com/Releases for more information. Looking to upgrade from an EOL release? See https://help.ubuntu.com/community/EOLUpgrades | 02:49 |
| stoned | The problem is that I am stuck on a live server, 10.04 LTS, and I need a new apache version (2.4) | 02:49 |
| stoned | Any current solution that will suffice. | 02:50 |
| stoned | hmm | 03:10 |
| stoned | I have an idea | 03:10 |
| stoned | setup a 64bit 14 lts chroot, setup apache 2.4/php/fastcgi there | 03:10 |
| stoned | let's see | 03:10 |
| Macer | you cant use lxc? | 03:15 |
| Macer | or vbox? :) | 03:15 |
| dw1 | I installed sendmail and set confAUTH_MECHANISMS and TRUST_AUTH_MECH in sendmail.mc and ran sendmailconfig to regenerate sendmail.cf and added a user but still can't seem to authenticate for relaying. What else do I need to do? Install saslauthd, perhaps? | 03:38 |
| stoned | Macer, who me? | 03:47 |
| stoned | no | 03:47 |
| stoned | this is a live server, rackspace, ubuntu 10 lts | 03:47 |
| stoned | I think I'll just debootsrap a 64bit debian stable chroot | 03:48 |
| stoned | I've never attempted to debootstrap ubuntu in my entire life | 03:48 |
| stoned | anyway, I guess I could 'jail' 2.4 apache inside a 64bit chroot | 03:48 |
| stoned | make sure it's working on 8080 w/ fast cgi, php5, apache 2.4 | 03:48 |
| stoned | then just disable the older httpd from 10.04 lts host | 03:49 |
| dw1 | got it. had to install sasl2-bin which gave me saslauthd and saslpasswd2 where I didnt know I had to add the user ;) | 03:54 |
| === sz0` is now known as sz0 | ||
| === sz0 is now known as sz0` | ||
| MACscr | stoned: or just get a second server that you upgrade and migrate things to and then retire the other one? | 04:21 |
| stoned | [22:46:47] <stoned> this is a live server | 04:21 |
| MACscr | aka, like the proper way to do it =P | 04:21 |
| stoned | if I could .. I would. | 04:21 |
| MACscr | and why cant you? | 04:22 |
| stoned | too many reasons I can't get into right now. | 04:22 |
| MACscr | well the hackish solutions you are looking for sound much worse | 04:22 |
| MACscr | especially since its a live system | 04:22 |
| dw1 | actually i was wrong didnt need saslpasswd2 just saslauthd | 04:23 |
| stoned | linux chroots are a hack. | 04:23 |
| stoned | Ever used BSD jails? | 04:23 |
| stoned | And yes, chrooting various server services in linux is also rather advisable. | 04:23 |
| stoned | Sounds to me you're unfamiliar w/ this approach of service segregation. | 04:24 |
| stoned | not to mention, even if your webserver gets hacked, it's still chrooted. It's a somewhat of a more secure setup as well. | 04:24 |
| stoned | Maybe I'll switch to nginx/fastcgi/php instead. | 04:26 |
| stoned | it's mostly a static site. | 04:26 |
| stoned | just high traffic. | 04:26 |
| stoned | actually | 04:27 |
| stoned | would you be willing to assist me in setting up fastcgi/mpm-worker/php5-cgi in apache 2.2 on 10.04? | 04:27 |
| stoned | That's the one thing I'm having issues. If I could switch to that, problems solved. | 04:27 |
| stoned | Currently the server is running Ubuntu 10.04 lts, apache 2.2 w/ mod_php(5) and apache mpm prefork | 04:28 |
| stoned | I make a clone of this server at this point | 04:28 |
| stoned | I then make changes to the clone (not the live server) and I successfully get fcgid/php5/cgi/apache 2.2 to work correctly | 04:29 |
| stoned | I then test my changes on the clone and it works | 04:29 |
| stoned | Then I do the exact same steps on the live server that i did on its clone (in theory it should work) | 04:29 |
| stoned | but I get internal server error. I can't find anything in /var/log/apache2/error.log (all error log directives point to this file) | 04:29 |
| stoned | So that's my current issue really. | 04:30 |
| stoned | If anyone of you could help me troubleshoot fcgid and the internal server error, I'd be very grateful | 04:30 |
| stoned | https://library.linode.com/web-servers/nginx/php-fastcgi/ubuntu-10.04-lucid hmm. | 04:40 |
| stoned | I wonder | 04:40 |
| sarkis | hey all, i added an upstart script and run initctl reload-configuration but i don't see my new script when i run initctl list | 05:05 |
| sarkis | weird thing is if i run service service-name start it is working | 05:05 |
| === justizin_ is now known as justizin | ||
| === jml_ is now known as jml | ||
| Siebjee | Does any one know a good cronjob scheduler like application that is scalable (like rundeck or jenkins) but is not one of those 2 ? | 07:53 |
| === sz0` is now known as sz0 | ||
| === sz0 is now known as sz0` | ||
| hxm | months ago I configured the port 25 to be opened for only localhost | 10:19 |
| hxm | but I dont remember where I did configure that | 10:19 |
| hxm | in /etc/postfix not | 10:19 |
| hxm | i think I put 127.0.0.1 in the config but dont remember where | 10:20 |
| mardraum | for postfix? in main.cf | 10:40 |
| nevercast | I wish to partition RHEL and install Ubuntu Server, I then wish to recover the partitions in use by RHEL for use by Ubuntu Server and configure grub for only one operating system. I also need to do this without ever being without an operating system, as I do not have the ability to use recover from removable media | 11:32 |
| nevercast | Suggestions ? | 11:32 |
| vfw | So you are going to create a dual-boot system? | 11:33 |
| nevercast | Only temporarily until Ubuntu is configured to work in place of RHEL, at which point I would like to remove RHEL | 11:34 |
| vfw | nevercast: Use non-destructive partition manager to shrink existing partitions in order to leave free space for the new install. | 11:34 |
| pmatulis | morning | 11:35 |
| vfw | nevercast: http://gparted.org/livecd.php | 11:35 |
| vfw | nevercast: ^^^^^^^^^^^^^ ... one example. | 11:38 |
| nevercast | vfw, that's for your guidence. It looks to be impossible anyway. As it is an OpenVZ VPS, and I do not have the typical block devices available | 11:38 |
| nevercast | *thanks | 11:38 |
| vfw | nevercast: Oh, a VPS. Well, that is different. | 11:39 |
| nevercast | I'm uncertain how to accomplish even a simple repartition in an OpenVZ Guest | 11:39 |
| vfw | nevercast: You would just need to re-install. | 11:39 |
| vfw | nevercast: Or get another VPS | 11:40 |
| nevercast | Unfortunately for myself, they only offer RHEL | 11:40 |
| vfw | nevercast: How about VBOX? | 11:40 |
| nevercast | VM in a VM.. Not a terrible idea, bit more overhead | 11:41 |
| nevercast | vfw, as I am only allocated a single IP address, I might run in to some concerns. Though I guess I could run the networking in bridged mode and disable most host networking services | 11:43 |
| soren | smoser: Before I start digging on my own, I was wondering if you have any idea where the time is spent when booting a cloud instance? How much is kernel, how much is initramfs, how much is fetching data from metadata service, how much is <each part of cloud-init>, etc.? | 11:45 |
| soren | smoser: I just saw http://t.co/JqAg7vOqgv this morning and wondered what those numbers really should be. 75,000 instances across 380 hosts is 197 per host. 6.5 hours is 390 minutes. That's ~two minutes per VM. 30 seconds per VM should be *plenty*. | 11:56 |
| soren | smoser: Also, it should be able to process a VM per core. A couple of million VM's should be doable in that timeframe on that many servers. | 11:59 |
| jamespage | soren, 2 minutes per vm? | 12:04 |
| soren | jamespage: Is my arithmetic off? | 12:04 |
| jamespage | soren, let me think | 12:05 |
| jamespage | soren, I had some stats on vm creation rate as well | 12:05 |
| soren | jamespage: This is assuming each node only handles on VM at a time. | 12:05 |
| soren | one VM, I mean. | 12:05 |
| jamespage | soren, at peak we hit 4.5 instances per second being created | 12:06 |
| soren | Then what the ¤!"#¤ is the rest of the time wasted on? | 12:06 |
| jamespage | soren, but that did drop off as load on the compute nodes increased | 12:06 |
| soren | jamespage: You shut VM's off again, right? | 12:06 |
| jamespage | soren, the cloud controllers (all three of them) get insanely busy | 12:06 |
| jamespage | soren, nope | 12:06 |
| soren | jamespage: Or did you leave all 11,000 running. | 12:06 |
| jamespage | soren, 75k running instances | 12:06 |
| soren | Sorry, 75k. | 12:06 |
| soren | Ok. | 12:07 |
| jamespage | soren, I was surprised exactly how insanely busy they got | 12:07 |
| soren | jamespage: Ok, so it's 200 VM's per host running at the same time. | 12:07 |
| jamespage | yup | 12:08 |
| soren | jamespage: How many cores? | 12:08 |
| jamespage | soren, 4 cores | 12:08 |
| soren | jamespage: Ok. | 12:09 |
| * soren is still not blown away | 12:09 | |
| soren | Sure, there's a lot of zeros at the end of that number, but one more should be possible. | 12:09 |
| jamespage | soren, to me it feels extremely intensive when setting up the instance | 12:09 |
| jamespage | prior to it actually being started | 12:09 |
| soren | It shoulnd't be. | 12:10 |
| soren | There's not much work to do. | 12:10 |
| soren | Copy a small image, start it up. After 10-20 seconds, it's idling. | 12:10 |
| jamespage | soren, that bit it quite quick | 12:10 |
| soren | So where's the bottleneck? | 12:10 |
| jamespage | soren, I'm talking about all of the bit prior to that happening | 12:10 |
| soren | Oh, OpenStack itself? Scheduling and whatnot? | 12:11 |
| jamespage | yup | 12:11 |
| soren | Yeah, *that* does not surprise me. | 12:11 |
| soren | But it shouldn't be that way. | 12:11 |
| soren | We should have no problem scheduling a couple of million instances in a few hours. | 12:11 |
| jamespage | soren, we actually started off running neutron networking - but the rabbitmq load was so high with both nova and neutron messages it started to reset connections | 12:12 |
| jamespage | soren, had to drop back to nova-network | 12:12 |
| soren | Also not surprising :) | 12:12 |
| * soren has to run | 12:12 | |
| jamespage | soren, lol | 12:13 |
| jamespage | later | 12:13 |
| hxm | i keep receiving this messages in my ssh session fatal: Read from socket failed: Connection reset by peer [preauth] | 12:21 |
| === sz0` is now known as sz0 | ||
| zul | jamespage: i thought i would change the channel topic since you know...quantal has reached eol ;) | 13:08 |
| === sz0 is now known as sz0` | ||
| jamespage | zul, for that we need a channel operator | 13:27 |
| zul | ah shoot | 13:27 |
| === sz0` is now known as sz0 | ||
| === sz0 is now known as sz0` | ||
| === sz0` is now known as sz0 | ||
| === Ursinha is now known as Ursinha-afk | ||
| === medberry is now known as Guest19303 | ||
| === Ursinha-afk is now known as Ursinha | ||
| tych0 | rbasak: have you used uvtool on the trusty i386 images? | 15:18 |
| tych0 | i don't think it's picking up my nocloud data | 15:18 |
| arosales | jamespage: smoser: apologies I didn't send last weeks meeting. Althought it was a quick one I didn't get minutes out so it looks like I will be chairing this week too. | 15:19 |
| rbasak | tych0: I think I've successfully used i386 images before. Not sure about the Trusty ones specifically. | 15:21 |
| tych0 | ok | 15:21 |
| tych0 | any thoughts on debugging? | 15:21 |
| dasjoe | Ah, people working on uvtool :) I asked this before, maybe you can answer this: "Hi, VMs created with uvt-kvm are not persistent, correct? What are possible use cases for uvtool, is there a way to make a VM persistent?" | 15:23 |
| rbasak | tych0: "virsh console <name>" just after starting the instance, and you should see console output. Then kill the instance, and examine logs on the disk using mount-image-callback on the file in /var/lib/uvtool/libvirt/images | 15:23 |
| rbasak | dasjoe: they are persistent. | 15:23 |
| rbasak | dasjoe: you may want them to automatically start. I can't remember if uvt-kvm does that or not, but you can adjust the autostart flag with virsh. | 15:24 |
| dasjoe | rbasak: Oh, okay. Thanks! I assumed they got undefined when stopped | 15:25 |
| rbasak | tych0: I've just successfully created a trusty i386 VM using uvtool 0~bzr92-0ubuntu1 | 15:27 |
| tych0 | huh | 15:27 |
| tych0 | oh | 15:27 |
| tych0 | i'm not on trusty :-) | 15:27 |
| tych0 | or rather | 15:27 |
| tych0 | my host isn't trusty | 15:27 |
| tych0 | that is proably part of it | 15:27 |
| rbasak | I can't think of any reason why it wouldn't work. | 15:28 |
| rbasak | The PPA (ppa:uvtool-dev/trunk) should have a recent version built for you. | 15:28 |
| tych0 | ok | 15:28 |
| tych0 | i should really just upgrade to trusty | 15:28 |
| === jrgifford__ is now known as jrgifford | ||
| zul | hallyn: ill try to get to libvirt this week | 15:42 |
| hallyn | zul: cool, it's starting to feel far behind :) | 15:42 |
| hallyn | thanks | 15:42 |
| zul | hallyn: or i can wait til 1.8.4 next week | 15:43 |
| zul | or whatever it is | 15:43 |
| hallyn | 9.7.h3 ? | 15:43 |
| smoser | arosales, no one ins going to complain if you say you're chairing again. :) | 15:48 |
| arosales | smoser: perhaps this time I'll remember to send the minutes. | 15:49 |
| www2 | hi can any one know why i get this error May 20 18:15:50 localhost postfix/smtpd[22424]: fatal: no SASL authentication mechanisms when i want send a test mail to my server? | 16:23 |
| www2 | the email send form my gmail adress | 16:23 |
| blu3ski3s | sounds like you need to setup SASL - http://www.postfix.org/SASL_README.html | 16:24 |
| patdk-wk | not really | 16:29 |
| patdk-wk | that says sasl is setup right | 16:29 |
| patdk-wk | but his sasl backend doesn't support whatever auth method he is using | 16:29 |
| hallyn | zul: feel like doing 1319717 while you package new libvirt? :) | 17:11 |
| hallyn | i betcha rbasak will give you a smooch if you do | 17:11 |
| hallyn | wait ,i mean he wont' do that | 17:12 |
| zul | no smooching | 17:12 |
| mdeslaur | hallyn: try bacon | 17:14 |
| hallyn | but not canadian bacon. i hear that'll set off a bad reaction | 17:15 |
| qhartman | I have a machine running 14.04 that I'm migrating a 12.04 apache config onto, and it's weirdly not writing a pid file. | 17:15 |
| qhartman | Even though it's starting correctly, the init script bombs because it thinks it's not | 17:15 |
| qhartman | the permissions on /var/run/apache2 seem fine, and the pid is set to write there | 17:16 |
| qhartman | but it never appears, even though I end up with running apache2 processes that are correctly listening. Any ideas? | 17:16 |
| qhartman | apache2ctl configtest seems to think everything is ok | 17:18 |
| lordievader | qhartman: Might be the change from 2.2 to 2.4. | 17:18 |
| rbasak | qhartman: check /var/log/kern.log for any apparmor denials. Is the pidfile in the standard location for Trusty? | 17:19 |
| qhartman | rbasak, don't see any apparmor entries, and I believe the pid is the standard place (/var/run/apache2) | 17:20 |
| shauno | I appear to have it in /run/apache2/apache2.pid on mine? | 17:21 |
| sarthor | HI, My ubuntu stop on grub and must need to press "keyboard ENTER" button, How can i fix this issue? | 17:21 |
| qhartman | shauno, /var/run is a symlink to /run | 17:22 |
| qhartman | so they should be the same in effect (unless apparmor is getting in the way and not mentioning it) | 17:22 |
| qhartman | lordievader, I'm sure that's the root of the problem at a high level, but I need to figure out something more specific to make it go | 17:23 |
| shauno | ah, I didn't spot that. seems you are looking in the right place then :) | 17:23 |
| qhartman | heh | 17:26 |
| qhartman | so, setting "PidFile ${APACHE_PID_FILE}" in apache2 conf allows it to work | 17:27 |
| qhartman | even though $APACHE_PID_FILE is set to "/var/run/apache2/apache2.pid" | 17:27 |
| qhartman | and that is what it was manually set to in the conf | 17:28 |
| qhartman | that's odd, but ok | 17:28 |
| [[lutchy]] | qhartman, What's odd? | 17:28 |
| qhartman | setting the the PidFile directive to use the variable, instead of the string that is the same as the variable contents, allowed it to start correctly | 17:29 |
| qhartman | was: PidFile /var/run/apache2/apache2.pid | 17:29 |
| qhartman | changed to: PidFile ${APACHE_PID_FILE} | 17:29 |
| [[lutchy]] | On that topic, first time working with Ubuntu (has a few Ubuntu Servers), I find that postfix in ubuntu would remove bounce messages from queue without re-trying or returning to sender? Is this a behavior that has to do with Relay messages? | 17:30 |
| qhartman | echo $APACHE_PID_FILE | 17:30 |
| qhartman | /var/run/apache2/apache2.pid | 17:30 |
| [[lutchy]] | Is $APACHE_PID_FILE an Environment variable ? | 17:31 |
| qhartman | [[lutchy]], yes | 17:31 |
| qhartman | Gets sourced in the from the "envvars" config in /etc/apache2 by the init script | 17:31 |
| [[lutchy]] | Uh, if I am understanding this correctly, you want the init script to modify Apache configs during to startup to applye the Pidfile directive ? | 17:32 |
| qhartman | no,m the standard config just uses the vars set in envvars so config changes that are global like that can be done in one place | 17:33 |
| qhartman | I'm just puzzled at why setting it to the same value as the var causes it not to work. I must be missing something, like it's not actually the same value and I'm somehow misreading it or something | 17:34 |
| qhartman | but, it's working now, so I'm not going to argue, and setting it to ${APACHE_PID_FILE} is probably more correct anyway | 17:35 |
| [[lutchy]] | Your explanation is a bit 'garbled' for me to quite understand your problem | 17:39 |
| qhartman | don't worry about it, it's under control now | 17:40 |
| qhartman | If anyone else is doing a 12.04 -> 14.04 server upgrade that involves apache2, this doc is helpful: http://httpd.apache.org/docs/2.4/upgrading.html | 18:00 |
| sarnold | qhartman: funny enough, I -think- I saw exact opposite issues on some of my VMs -- something like the init script didn't report any errors but the server wasn't running when the init script finished. | 18:02 |
| sarnold | qhartman: of course my apache configs in those VMs were pretty horrible at that point, I blamed it on the frankenmess I had created :) | 18:02 |
| patdk-wk | haven't had any issues :) | 18:03 |
| sarthor | HI, My ubuntu stop on grub and must need to press "keyboard ENTER" button, How can i fix this issue? | 18:25 |
| patdk-wk | GRUB_RECORDFAIL_TIMEOUT=5 | 18:28 |
| === soren changed the topic of #ubuntu-server to: Ubuntu Server discussion and support | For general (not server specific) support visit #ubuntu | IRC Guidelines: https://wiki.ubuntu.com/IrcGuidelines | Get involved: https://wiki.ubuntu.com/ServerTeam/GettingInvolved | Docs and resources: https://help.ubuntu.com/12.04/serverguide/ | Utopic (12.10) roadmap bugs, http://status.qa.ubuntu.com/reports/ubuntu-server/release-bugs.html | ||
| === soren changed the topic of #ubuntu-server to: Ubuntu Server discussion and support | For general (not server specific) support visit #ubuntu | IRC Guidelines: https://wiki.ubuntu.com/IrcGuidelines | Get involved: https://wiki.ubuntu.com/ServerTeam/GettingInvolved | Docs and resources: https://help.ubuntu.com/12.04/serverguide/ | Utopic (14.10) roadmap bugs, http://status.qa.ubuntu.com/reports/ubuntu-server/release-bugs.html | ||
| soren | zul: There you go. | 19:02 |
| soren | Hm.. | 19:02 |
| zul | thanks | 19:02 |
| === soren changed the topic of #ubuntu-server to: Ubuntu Server discussion and support | For general (not server specific) support visit #ubuntu | IRC Guidelines: https://wiki.ubuntu.com/IrcGuidelines | Get involved: https://wiki.ubuntu.com/ServerTeam/GettingInvolved | Docs and resources: https://help.ubuntu.com/14.04/serverguide/ | Utopic (14.10) roadmap bugs, http://reports.qa.ubuntu.com/reports/rls-mgr/rls-r-tracking-bug-tasks.html#s | ||
| soren | There. Even better. | 19:03 |
| soren | Gah, except it's too long. | 19:03 |
| sarnold | you could probably trim the "Get involved: " text, it's in the url too | 19:03 |
| === soren changed the topic of #ubuntu-server to: Ubuntu Server discussion and support | For general (not server specific) support, try #ubuntu | IRC Guidelines: https://wiki.ubuntu.com/IrcGuidelines | https://wiki.ubuntu.com/ServerTeam/GettingInvolved | Docs and resources: https://help.ubuntu.com/14.04/serverguide/ | Utopic (14.10) roadmap bugs, http://reports.qa.ubuntu.com/reports/rls-mgr/rls-r-tracking-bug-tasks.html#server | ||
| soren | sarnold: Good idea. Done. | 19:03 |
| soren | booyah | 19:03 |
| soren | Only took 4 revisions. | 19:04 |
| sarnold | not bad :) hehe | 19:04 |
| qhartman | sarnold, heh, I have definitely been in frankenmess land with apache. At my last job I inherited one config that was over 3k lines (not including comments!) that was all in one massive file | 19:07 |
| qhartman | Oh the tears, they were bitter | 19:08 |
| qhartman | apparently it was something that had been getting rolled forward in this company since apache 1.x days | 19:08 |
| sarnold | qhartman: yikes. the pain. | 19:08 |
| qhartman | The one upside with that setup is there was one place to look to figure out what was wrong | 19:09 |
| lordievader | Wow, 3k lines. Purge and rebuild? | 19:10 |
| qhartman | lordievader, yeah, I ended up setting up a temporary box and migrated every site by hand over to it individually to recreate a sane config, and then copied the config back to the production box | 19:11 |
| lordievader | That is what I like about Puppet, one place to configure all the servers/virtual-machines. | 19:11 |
| qhartman | it took the better part of a week to tease all the pieces apart | 19:11 |
| qhartman | yeah, this was back when your configuration configuration management choices were cfengine or SVN. | 19:12 |
| jamespage | med_, seriously - bug 1320960 | 19:12 |
| jamespage | ? | 19:12 |
| uvirtbot | Launchpad bug 1320960 in cloud-archive "Add Trusty Icehouse pkgs to UCA" [Undecided,New] https://launchpad.net/bugs/1320960 | 19:12 |
| med_ | jamespage, several folks I work with expected that type of behavior to be the default.... | 19:13 |
| * med_ explained that was not the default | 19:13 | |
| jamespage | med_, hmmmm | 19:14 |
| med_ | jamespage, feel free to close with emphasis. | 19:14 |
| jamespage | I'll think about it | 19:14 |
| jamespage | we might be able todo something transparent | 19:14 |
| med_ | an empty repo might do the trick | 19:14 |
| jamespage | so it works but is actually a no-op | 19:14 |
| keithzg | Ouch, my btrfs drive appears to be so screwed that it's going "Check tree block failed, want=1266349326848, have=18350744064220987392" and refusing to mount. Guess I learn about fixing/recovering btrfs now . . . | 19:30 |
| wam | kees: you got this i hope? https://btrfs.wiki.kernel.org/index.php/Problem_FAQ | 19:43 |
| wam | keithzg_: even | 19:43 |
| keithzg_ | wam: Yeah, I'm pretty sure it isn't a "bug" per se though, the power in my building has been very wonky lately and apparently we can't afford more UPS' so although all the primary servers other folks use are on them, I don't have one for my testbed/personal one at the office, so it's been going down hard from time to time :( | 19:47 |
| wam | keithzg_: have fun | 19:47 |
| axisys | why would swap in use when 25G memory not in use? | 19:55 |
| axisys | Mem: 32909200k total, 7345056k used, 25564144k free, 181956k buffers | 19:55 |
| axisys | Swap: 499708k total, 290772k used, 208936k free, 2810712k cached | 19:55 |
| axisys | should I change the swappiness to 0 ? currently it is 10 | 19:55 |
| axisys | and do a swapoff/swapon ? | 19:55 |
| sarnold | axisys: it's probably not worth the effort | 19:57 |
| axisys | xymon alerting 95% swap in use.. even though tons of free ram.. what do you recommend? | 19:57 |
| axisys | sarnold: ^ | 19:57 |
| sarnold | axisys: it might add another four or five seconds to shutdown to 'swapoff' the thing. otherwise it is probably not going to influence much of anything. | 19:57 |
| wam | axisys: always use swappiness 0. Nothing else is sane if you have memory. Linux will then first use most of the memory. | 19:58 |
| axisys | wam: ok | 19:58 |
| wam | axisys: and do a swapoff / swapon cycle after setting it. linux won't clean up otherwise. | 19:59 |
| axisys | wam: just did that.. swap use is now 0 | 19:59 |
| wam | axisys: but you need to know - the used swap in your server is even cached, so not much penalty from it. | 19:59 |
| wam | axisys: it's just that linux could just throw the cache away as soon as there's a bottleneck | 20:00 |
| wam | axisys: and it's the bottleneck detection that might not be what you want. | 20:00 |
| axisys | so I really should not generate alert when swap usage is over 95% | 20:01 |
| wam | axisys: let's say 50% swap usage doesn't say much about your memory. | 20:01 |
| axisys | since I did not see any slowness with 25 gig ram around | 20:01 |
| wam | right, it's all cached. | 20:01 |
| wam | it's just cold pages that could be thrown away immediately and that would not have to be written to disk first. | 20:01 |
| [[lutchy]] | If you have 25GB of RAM | 20:02 |
| wam | here, cold pages are not THAT cold. So swap always has to keep up. So we do swappiness=0. It's with occupied virtual machines. | 20:02 |
| === keithzg_ is now known as keithzg | ||
| [[lutchy]] | With only ONE OS | 20:02 |
| axisys | [[lutchy]]: I have 32G total | 20:02 |
| [[lutchy]] | Turn of swap | 20:03 |
| axisys | [[lutchy]]: yes | 20:03 |
| wam | never turn off swap | 20:03 |
| wam | you will have NO time to fix oom | 20:03 |
| [[lutchy]] | Yes, Turn off swap | 20:03 |
| wam | and oom is ALWAYS wrong. Per definition | 20:03 |
| [[lutchy]] | The only you have to consider if you an programs that need more than 32GB | 20:03 |
| [[lutchy]] | s/only/only issue/ | 20:03 |
| axisys | currently swap size is 512MB /dev/mapper/volg0-swap | 20:03 |
| wam | even 2-4 GB swap is ok. The more the better. Because you earn TIME when oom is active. | 20:04 |
| wam | except that a hard reset on every memory shortcoming is acceptable. | 20:04 |
| [[lutchy]] | Swap is a safety net but if you have programs running out of memory --- Which will be indicated, then consider swap | 20:04 |
| [[lutchy]] | Having Swap on, the Kernel have consider Swap while doing Memory management | 20:05 |
| wam | It's all very complicated ;) | 20:05 |
| sarnold | axisys: much better would be to alert on swap in and swap out traffic | 20:06 |
| wam | correct ^^ | 20:06 |
| sarnold | axisys: check the 'vmstat 1' output, read the 'si' and 'so' columns | 20:06 |
| wam | this will hurt i/o performance. | 20:06 |
| sarnold | axisys: most of the time you shouldn't have any; a little is not too bad. a lot is bad. | 20:06 |
| axisys | yep they are 0 | 20:06 |
| sarnold | axisys: I don't know any good way to alert on high values, but it is a more reliable indicator of trouble than a lot of other metrics | 20:07 |
| [[lutchy]] | I really depends on what he/she is running | 20:09 |
| axisys | he | 20:09 |
| axisys | this system has tons of scripts running and backup client.. thats all | 20:10 |
| [[lutchy]] | I can't see, one OS, with few functions having a lot of 'swap in' and swap out' | 20:10 |
| axisys | nothing bad | 20:10 |
| sarnold | still, if you're using much swap it's probably time to buy more memory :) | 20:10 |
| [[lutchy]] | Unless there is a memory leak | 20:10 |
| [[lutchy]] | sarnold, He/she says he has 32GB | 20:10 |
| sarnold | (and by 'using swap' i don't mean your swap space is in use; I mean the 'si' and 'so' columns have much use) | 20:10 |
| axisys | sarnold: no.. based on our metrics graph..only some app probably behaved bad for few minutes | 20:11 |
| sarnold | [[lutchy]]: good thing he wasn't using much swap :) | 20:11 |
| axisys | lot of rams | 20:11 |
| sarnold | [[lutchy]]: it gets expensive to add another 32 or 64, hehe | 20:11 |
| axisys | [[lutchy]]: he :-) | 20:11 |
| [[lutchy]] | I have only 6GB on my server | 20:11 |
| [[lutchy]] | Which is Windows, then Hyper V clients | 20:12 |
| [[lutchy]] | I have two Hype V clients running .. I have -- mild swap (after a few Windows Updates) | 20:13 |
| Havenstance | is there an apt-get solution for ntopng on 13.10? | 20:13 |
| [[lutchy]] | Both are Ubuntu Servers | 20:13 |
| axisys | understand we don't have any slowness.. so disabling alert about swap measure is probably best bet | 20:13 |
| [[lutchy]] | Well, I think, you seem not to understand, I would say best bet is to have swap on | 20:14 |
| sarnold | but there's probably no point in alerting anyone about high swap allocation | 20:14 |
| dw1 | Havenstance: do-release-upgrade :) | 20:14 |
| [[lutchy]] | Then look how at 'Swap in' and 'Swap out' account to what what sarnold mentioned | 20:14 |
| [[lutchy]] | right ... | 20:15 |
| dw1 | Havenstance: oh you have til July nm | 20:15 |
| [[lutchy]] | That's probably the KERNEL on Heavy IO | 20:15 |
| axisys | [[lutchy]]: swap is on | 20:15 |
| Havenstance | dw1 just waiting for the newest release of zentyal to drop on 14.04 then i'll update it | 20:15 |
| [[lutchy]] | Is your DISK slow ? | 20:15 |
| dw1 | Havenstance: https://raymii.org/s/tutorials/ntop-ng-installation-on-Ubuntu.html -- use checkinstall to create a package | 20:16 |
| [[lutchy]] | 70MB is slow in a Server environment ... for me at least | 20:16 |
| Havenstance | thanks man | 20:16 |
| axisys | [[lutchy]]: how do I check if disk is slow? hdparm ? | 20:17 |
| [[lutchy]] | Harddrive can be handicap.. I am not familiar with raid setup, which can improve read/write performance | 20:18 |
| [[lutchy]] | Without OS Write Cache, my Hard Drive can only musta 9/MBs on write | 20:19 |
| [[lutchy]] | With OS Cache, I get about 70MB write cache | 20:20 |
| axisys | 157.65 MB/sec read | 20:21 |
| [[lutchy]] | Looks like you have SSD drive | 20:21 |
| [[lutchy]] | ? | 20:23 |
| axisys | I dont think so.. how do I find out? | 20:23 |
| axisys | HP DL360p | 20:23 |
| axisys | smartctl does not say since behind raid controller | 20:24 |
| [[lutchy]] | With 157.xxx, it's most likely SSD drive | 20:25 |
| [[lutchy]] | Ubuntu has lshw | 20:25 |
| axisys | running | 20:25 |
| axisys | *-disk:1 | 20:26 |
| axisys | description: SCSI Disk | 20:26 |
| axisys | product: LOGICAL VOLUME | 20:26 |
| axisys | vendor: HP | 20:26 |
| fr36 | hello | 20:29 |
| axisys | [ 36.786112] hpsa 0000:02:00.0: RAID device c2b0t0l0 added. | 20:31 |
| axisys | [ 36.786319] scsi 2:0:0:0: RAID HP P420i 3.04 PQ: 0 ANSI: 5 | 20:31 |
| fr36 | can anyone help me? | 20:31 |
| wam | fr36: probably not | 20:48 |
| fr36 | why not? what is this channel for? | 20:48 |
| RoyK | !ask | fr36 | 20:49 |
| ubottu | fr36: Please don't ask to ask a question, simply ask the question (all on ONE line and in the channel, so that others can read and follow it easily). If anyone knows the answer they will most likely reply. :-) See also !patience | 20:49 |
| wam | fr36: for specific information or questions about ubuntu server somehow i guess | 20:49 |
| RoyK | wam: better give him the bot than asking him implicitly to leave :P | 20:50 |
| wam | I didn't want him to leave. | 20:50 |
| RoyK | well, he didn't | 20:50 |
| wam | RoyK: though you're right of course. Didn't know the bot | 20:50 |
| wam | ubottu: hello | 20:50 |
| fr36 | I need to configure my ubuntu box to authenticate to an ActiveDirectory | 20:50 |
| RoyK | fr36: login or samba or what? | 20:51 |
| wam | fr36: the box or the users on the box? | 20:51 |
| fr36 | the users on the box | 20:51 |
| fr36 | my ubuntu box will be a client | 20:52 |
| fr36 | I thought I should be using kerberos | 20:52 |
| RoyK | so users login to ubuntu and want their profile/homedir from a windows share? | 20:52 |
| wam | fr36: http://www.lmgtfy.com/?q=ubuntu+user+login+active+directory | 20:53 |
| wam | fr36: or is there anything specific? | 20:53 |
| fr36 | basically users need to authenticate to a AD domain to be able to use the mail server | 20:54 |
| wam | fr36: mostly https://help.ubuntu.com/community/ActiveDirectoryWinbindHowto | 20:54 |
| wam | fr36: and yes, kerberos is good. | 20:54 |
| fr36 | my specific question is: how can I find the KDC server? | 20:55 |
| fr36 | is there a way to locate it? | 20:55 |
| RoyK | fr36: it's the dc | 20:55 |
| RoyK | fr36: the domain controller | 20:55 |
| fr36 | right | 20:56 |
| RoyK | fr36: the domain controller(s) | 20:56 |
| fr36 | yes it is, but I have to know its name | 20:57 |
| RoyK | fr36: ask the sysadmin | 20:57 |
| RoyK | fr36: that's just a machine in the local network | 20:57 |
| fr36 | that's not an option, it's a big company | 20:58 |
| fr36 | however the KDC is the Key Distribution Center | 21:04 |
| fr36 | it's not the Domain Controller | 21:04 |
| RoyK | fr36: it is | 21:05 |
| fr36 | RoyK: I need a way to locate it | 21:05 |
| RoyK | fr36: it'll be a *very* large company to have a dedicated KDC | 21:05 |
| fr36 | RoyK: it is | 21:06 |
| RoyK | how large? | 21:06 |
| fr36 | thousands of employees | 21:06 |
| RoyK | dig -t srv _kerberos._udp.XXXXXX.com | 21:06 |
| fr36 | RoyK: thanks | 21:08 |
| RoyK | didn't work here, though | 21:08 |
| fr36 | why not? | 21:09 |
| Armadillos | RoyK: While it's possible the DC and KDC are on the same server, most companies will have that seperate. Are you on a windows computer that is currently logged into the domain? | 21:09 |
| fr36 | Armadillos: I can get a hold of one | 21:10 |
| fr36 | what should I do? | 21:10 |
| Armadillos | fr36: Go to that computer, and once your logged in, go to a command prompt. Type in this command: echo %LOGONSERVER% | 21:11 |
| Armadillos | That will give you a DC | 21:11 |
| fr36 | but not a KDC | 21:12 |
| * wam learns | 21:12 | |
| fr36 | I need a KDC to configure a kerberos client | 21:12 |
| wam | I'd just use wireshark and see where traffic goes ;) | 21:13 |
| Armadillos | KDC is a Key distribution system used to give out Windows Keys, not for Kerboros. | 21:13 |
| RoyK | Armadillos: I'm not the one asking questions. I work for a college with some 20k students and 2k employees. We have DC and KDC on the same servers. | 21:13 |
| Armadillos | RoyK: Yeah, sorry about that. :) | 21:13 |
| === broder_ is now known as broder | ||
| fr36 | Armadillos: https://help.ubuntu.com/community/Kerberos just says it needs | 21:14 |
| Armadillos | fr36: So the KDS they're talking about in that article is usually handled by a DC. It's part of the Domain Controller. | 21:15 |
| fr36 | I need to configure the Kerberos servers for my realm but I don't know their name | 21:17 |
| wam | fr36: you haven't told us if the dc works or not for this purpose. | 21:18 |
| Armadillos | fr36: Are you trying to build another Kerberos system, or trying to get the current server connected to an existing server? | 21:18 |
| fr36 | just trying to act as akerberos client | 21:19 |
| Armadillos | Then the current DCs in place should be fine. You just need to have that server "point" to one of the DCs. | 21:20 |
| fr36 | I just need to know what to write in the kdc and int hte admin_server properties of my realm | 21:20 |
| Armadillos | fr36: Run that command on a windows box, and it will give you a name of a DC that can handle the KDC requests. | 21:21 |
| fr36 | should then put that name in the kdc line? | 21:21 |
| wam | wtf | 21:22 |
| Armadillos | fr36: You'll probably want to use the FQDN, but yes. | 21:40 |
| fr36 | Armadillos: thank you | 21:42 |
| mikey85 | http://irclogs.ubuntu.com/2014/05/20/%23ubuntu.txt | 22:01 |
| mikey85 | look at those logs | 22:01 |
| mikey85 | I was accused wrongly | 22:01 |
| mikey85 | press F3 and type in mikey85 | 22:02 |
| mikey85 | You'll see | 22:02 |
| mikey85 | that's why I was banned from ubuntu | 22:02 |
| mikey85 | by ikonia | 22:02 |
| mikey85 | === andrex|on is now known as mikey85_ | 22:03 |
| mikey85 | see that? | 22:03 |
| mikey85 | that was in the log | 22:03 |
| mikey85 | ikonia has bullied me time and time again | 22:03 |
| Corey | mikey85: That's not topical here. | 22:03 |
| mikey85 | as well as weaseling her way out of dinner >:( | 22:03 |
| mikey85 | ahhh lol | 22:03 |
| mikey85 | i see lol | 22:04 |
| mikey85 | so she has no say in here? :D | 22:04 |
| === MACscr is now known as Captain_Obvious | ||
| === Captain_Obvious is now known as MACscr | ||
| phillw | Hi good people, I'm following http://maas.ubuntu.com/docs1.5/install.html#pkg-install and it seems very badly broken (the ppa doesn't work, the mass-dhcp and mass-dns do not exist. Is there a working set of instructions for installing cloud onto a trusty server? | 22:53 |
| sarnold | phillw: you may have better success in #maas | 22:54 |
| phillw | sarnold: thanks :) | 22:55 |
| sarnold | phillw: probably you can ignore the ppa advice.. | 22:55 |
| phillw | sarnold: that'd be fine, but the 'installer does not start... maybe a tasksel command, let me go see | 22:56 |
| === Lcawte is now known as Lcawte|Away | ||
| fridaynext | it looks like plex is no longer available to update via apt-get update / upgrade... | 23:40 |
| fridaynext | anyone else had this issue? | 23:40 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!