mwhudson | um | 00:10 |
---|---|---|
mwhudson | why might a cloud image end up with #MANUAL | 00:10 |
mwhudson | nameserver 0.0.0.0 | 00:10 |
mwhudson | in /etc/resolv.conf? | 00:10 |
mwhudson | i'm pretty sure dhcp is passing a nameserver | 00:10 |
mwhudson | (it's not actually running in a cloud) | 00:10 |
=== Lcawte is now known as Lcawte|Away | ||
TriJetScud | I'm looking for guidance on creating custom ubuntu cloud images... | 00:54 |
sarnold | TriJetScud: seen this yet? it might be enough to keep you from having to prepare you own customized images: http://cloudinit.readthedocs.org/en/latest/ | 00:56 |
TriJetScud | I've seen this but cloud-init doesn't support microsoft's system center yet | 00:57 |
TriJetScud | I'm just looking at how the ubuntu cloud images are created, scripts wise | 01:06 |
=== Nigel_ is now known as G | ||
=== phuh is now known as talves_ | ||
=== markthomas_ is now known as markthomas | ||
=== talves_ is now known as phuh | ||
shuggans | hi all - have zoneminder installed on ubuntu server 12.04 - trying to redirect its recording data (/usr/share/zoneminder/images) to /media/videostore/DVR/images using fstab but getting mount errors on boot | 05:21 |
shuggans | can anyone help me figure out what I'm doing wrong? | 05:22 |
larsemil | does anyone have any hints on how to troubleshoot ldap + tls? I dont get it to work. I followed this guide: https://help.ubuntu.com/12.10/serverguide/openldap-server.html#openldap-tls | 06:02 |
larsemil | when i run a ldapsearch i get: root@master:~# ldapsearch -x -ZZ -H ldap://83.209.243.253 -b dc=daladevelop,dc=se │····································································· | 06:02 |
larsemil | ldap_start_tls: Protocol error (2) │····································································· additional info: unsupported extended operation | 06:02 |
larsemil | sorry strange paste | 06:02 |
frt | hi folks | 07:02 |
frt | I get crontab permission denied error | 07:02 |
frt | when I execute vi sh in cli it does work | 07:03 |
larsemil | http://serverfault.com/questions/602258/ubuntu-14-04-openldap-tls-problems | 07:04 |
frt | I have typed crontab -u root -e | 07:04 |
frt | same command is there its all cool | 07:04 |
=== Lcawte|Away is now known as Lcawte | ||
=== Trey is now known as Guest20829 | ||
=== ndee_ is now known as ndee | ||
LeMike | hello. the "original" PHP version is always different from the one on ubuntu-servers because they have their own patches - so the Patch-Number is different. How near is it to the original one if it quotes "PHP 5.3.2-1ubuntu4.24" ? or which PHP version is "PHP 5.3.2-1ubuntu4.24" truly? | 07:49 |
TJ- | LeMike: check the package's Changelog | 07:51 |
=== Trey is now known as Guest35373 | ||
LeMike | TJ-: Where in the changelog do I see that "5.5.9+dfsg-1ubuntu3" is the same as "PHP 5.5.9" or "PHP 5.5.something" ? there are also the Ubuntu-Version-Numbers for PHP | 09:00 |
TJ- | LeMike: You look for the earliest 5.5.9 entry, and then work forward to the one you're interested in. Each version will add something in terms of additional patches - possibly backports from upstream, possibly local Debian/Ubuntu bug-fixes. But, essentially, the base-version of the package (before any -XubuntuY extension) is the PHP version of the upstream source | 09:18 |
caribou | jamespage: thanks for the sosreport SRU review; | 09:40 |
caribou | jamespage: I'm surprized about the DEP3 comment, I used quilt --dep3 to create the patch headers | 09:40 |
jamespage | caribou, np | 09:44 |
caribou | jamespage: what's wrong about it ? | 09:44 |
jamespage | caribou, here's one that uses the Description field correctly: | 09:45 |
jamespage | http://paste.ubuntu.com/7600696/ | 09:45 |
caribou | jamespage: ah, the spacing in the Desc ? | 09:46 |
LeMike | Ah! So the 5.5.3+dfsg is the one, that is in sync with the PHP 5.5.3, TJ- ? | 09:46 |
TJ- | LeMike: Correct | 09:46 |
LeMike | Thanks. That was easy TJ- :D | 09:46 |
jamespage | caribou, you need to use the Description field name | 09:46 |
jamespage | and indent to ensure it can be parsed as a single field | 09:47 |
TJ- | LeMike: "+dsfg" simply means that Debian modified the original upstream source package to conform with their definition of "free" software. see https://wiki.debian.org/DFSGLicenses | 09:48 |
caribou | jamespage: crap, I missed that one, will fix it | 09:49 |
caribou | jamespage: on that topic, is it good practice to add those DEP3 header to patches ? I see many packages that don't have them | 09:50 |
jamespage | caribou, its really good practice | 09:50 |
caribou | jamespage: ok will keep doing it, but correctly this time ;-) | 09:51 |
Codmadness | Can anyone help me with this error http://pastie.org/9263748 | 09:55 |
YamakasY | I can login to a 14.04 fresh install using console but SSH doesn't accept my password... what could be wrong ? | 09:57 |
sonne | when was 14.04 supposed to be released on LTS server again? | 11:00 |
sonne | should have been around these days... | 11:00 |
sonne | i only remember "june" though :) | 11:00 |
YamakasY | when I deploy a 12.04 ubuntu install I can login with the rootpw using ssh, the same deploy for 14.04 doesn't accept the password | 11:01 |
Codmadness | Can anyone help me with this error http://pastie.org/9263748 | 11:17 |
bgardner | Codmadness: Give me a minute to read up | 11:17 |
Codmadness | Ok | 11:17 |
bgardner | Codmadness: I saw you over in #ubuntu, did they already give you instructions? What have you already done? | 11:18 |
Codmadness | Well I've tried reinstalling, purging mysql, removing mysql and keep getting errors. Everytime I do apt-get install mysql-server it says 2 are not fully installed | 11:19 |
Codmadness | Which in the end ends up with E: Sub-process /usr/bin/dpkg returned an error code (1) | 11:20 |
bgardner | Codmadness: Ok, and is your other comment I see about 'port 3307 in use' related or not? | 11:20 |
Codmadness | Well I know that nothing else is 3307 | 11:21 |
Codmadness | I even changed to port to see a difference which nothing worked. | 11:21 |
Codmadness | #mysql Said E: Sub-process /usr/bin/dpkg returned an error code (1) is a ubuntu issue | 11:21 |
=== Lcawte is now known as Lcawte|Away | ||
bgardner | Oh, I see; initscript action start failed. That's probably the port in use. One moment. | 11:22 |
Codmadness | Yet I get E: Sub-process /usr/bin/dpkg returned an error code (1) everytime I reinstall or install mysql-server. How could it stop the install process? | 11:25 |
Codmadness | I also get that error everytime I upgrade | 11:25 |
bgardner | Codmadness: Not sure yet, could you get me a pastebin of this: sudo lsof -Pnl +M -i4 | 11:26 |
Codmadness | Do you mind me using pastie.org? Its much eaiser to use pastie.org on my pi | 11:27 |
bgardner | Sure | 11:28 |
Codmadness | http://pastie.org/9263980 | 11:28 |
bgardner | Codmadness: What's the output of lsb_release -a | 11:29 |
bgardner | Codmadness: And then after that, paste the output of: sudo netstat -tulpn | 11:32 |
Codmadness | http://pastie.org/9263991 | 11:33 |
bgardner | Codmadness: What's in the mysql logs? | 11:38 |
bgardner | Codmadness: Or syslog, related to mysql | 11:38 |
pmatulis | morning | 11:39 |
Codmadness | LEt me check | 11:40 |
Codmadness | http://pastie.org/9264027 | 11:43 |
Codmadness | Thats what happens when i try to start | 11:43 |
bgardner | Codmadness: While I look at that please paste the contents of /etc/init.d/mysql | 11:44 |
bgardner | Codmadness: Need to step away for a couple minutes but I'll be back. | 11:46 |
Codmadness | k | 11:52 |
bgardner | Codmadness: I'm back. This promises to take a while, let's move this conversation private so we don't clutter the channel. | 11:55 |
TJ- | Codmadness: I'd suggest first disabling the starting of mysql so you can complete the upgrade. Once you've done that, you can then work on it by manually starting the mysqld and investigating why processes seem to left running when they're told to stop | 11:59 |
Codmadness | TJ- Since I've installed it I have rebooted the vps so it should have killed it | 12:00 |
TJ- | Codmadness: But have you fixed the package issues caused? If not, the package will still be 'unconfigured' | 12:01 |
Codmadness | Well I've done --fix-broken | 12:02 |
TJ- | Codmadness: After restarting, did you check if the mysqld is running? | 12:02 |
Codmadness | TJ- Well I checked system monitor | 12:03 |
TJ- | Codmadness: check the PID file contents matches the running instance PID, with " PIDM=$(sudo cat /run/mysqld/mysqld.pid); ps -efly | grep $PIDM " | 12:06 |
Codmadness | The pid constantly changes | 12:06 |
TJ- | Codmadness: You should get 2 lines, one of which ends with "/usr/sbin/mysqld" | 12:06 |
TJ- | It does? That suggests the service is restarting constantly | 12:06 |
Codmadness | It grows up in numbers everytime | 12:07 |
Codmadness | Wait thats what aus does :/ | 12:08 |
TJ- | Codmadness: OK, I'd suggest disabling the mysql service for starters so you can work with it manually. You can override the upstart init job by doing "echo manual | sudo tee -a /etc/init/mysql.override" | 12:10 |
TJ- | Codmadness: If the service is repeatedly restarting, that might explain why the init script failed to keep track of the current PID, and therefore fails to stop it with "no such instance" | 12:11 |
Codmadness | Ok one second | 12:11 |
Codmadness | Ok let me run that one second | 12:14 |
Codmadness | Ok i ran the echo command | 12:15 |
TJ- | Codmadness: now you need to stop/kill any existing instance | 12:17 |
Codmadness | kill -mysql? | 12:17 |
TJ- | Codmadness: "sudo pkill mysqld" | 12:18 |
Codmadness | Ok done | 12:18 |
TJ- | Codmadness: After that, check for any remaining instances with "pgrep mysqld" | 12:18 |
Codmadness | It returns nothing... | 12:19 |
TJ- | Codmadness: good, that means no mysqld instances are now running | 12:19 |
Codmadness | Oh, I thought that was bad xD | 12:20 |
TJ- | Unix commands are generally quiet when they are successful. If something goes wrong, they complain | 12:21 |
TJ- | I'm reading the mysql upstart script to see if it could be doing anything special | 12:21 |
Codmadness | Oh, ok thanks | 12:22 |
TJ- | Codmadness: OK, nothing amazing there. It does some basic checks like ensuring disk space, and only starts mysqld if those pass. So the fact it starts tells us that part is OK. So, if mysqld is respawning (thus causing the PID to change) then something it finds when starting is the problem. So if you start it manually you might be able to catch some clues | 12:30 |
Codmadness | So I shall start mysql right? | 12:31 |
TJ- | Codmadness: Try "sudo /usr/sbin/mysqld" see if it exits with some errors, or continues running | 12:31 |
Codmadness | root@playmcpe:~# sudo /usr/sbin/mysqld | 12:33 |
Codmadness | 140606 16:36:34 [Warning] Using unique option prefix key_buffer instead of key_buffer_size is deprecated and will be removed in a future release. Please use the full name instead. | 12:33 |
Codmadness | root@playmcpe:~# | 12:33 |
Cyberspirit | command to see the size of a file from friggen bash console? | 12:33 |
Codmadness | Woops should have pastbinned it | 12:34 |
Codmadness | Didn't realize it was that big | 12:34 |
TJ- | Codmadness: Is that all you see? | 12:34 |
Codmadness | Yes TJ- | 12:34 |
TJ- | Codmadness: It hasn't returned to a shell command prompt? | 12:34 |
Codmadness | Nope | 12:35 |
Cyberspirit | command to see the size of a file from friggen bash console? | 12:35 |
TJ- | Codmadness: OK, you need to send it a SIGQUIT to exit it. Press Ctrl+\ and wait a few seconds, then it should exit | 12:35 |
Codmadness | SIGQUIT | 12:35 |
Codmadness | Woops | 12:35 |
Codmadness | Command not found | 12:36 |
Cyberspirit | stat | 12:36 |
TJ- | Codmadness: You have a shell prompt again? | 12:37 |
gry | Cyberspirit: "ls -lah /path/to/file" | 12:38 |
Codmadness | Well it says sigquit isn't a command | 12:38 |
TJ- | Codmadness: Haha! You don't type SIGQUIT ... that is the name of the 'signal' sent to the process when you press Ctrl+\ | 12:38 |
gry | Cyberspirit: "ls -la /path/to/file" in something raw, probably bytes | 12:39 |
TJ- | Codmadness: as long as you have the shell prompt back you're fine | 12:39 |
TJ- | Codmadness: As that didn't fail, I'm going to guess the problem was transient and has gone away. On that assumption remove the manual override with "sudo sed 's/manual//' /etc/init/mysql.override" and then restart the service with "sudo service mysql start" | 12:39 |
=== Trey is now known as Guest57159 | ||
Codmadness | Wait so when I run sudo /usr/sbin/mysqld it should popup a shell script? | 12:40 |
TJ- | Codmadness: No. In the interactive shell, doing "sudo /usr/sbin/mysqld" starts mysql server in the foreground so it doesn't return to the shell. Therefore, you have to tell the process to quit to get back to the shell. | 12:41 |
TJ- | Codmadness: The shell is what provides the "$USER@$HOST:~$" prompt where your command are interpreted | 12:42 |
Codmadness | Oh | 12:42 |
TJ- | Codmadness: So, as I said, remove that override and restart the service and then check the mysql log-file again and see if it is running correctly | 12:43 |
xtriz | i have been using gentoo on my machine, and tired due to user intervention in maintaining system, is ubuntu stable enough that i do not need to spend the time in maintaining it ? | 12:43 |
Codmadness | ok | 12:43 |
xtriz | right now i am under tight time constraint and can't afford to waste the time in fixing system | 12:43 |
ikonia | xtriz: it's in use in the enterprise, it's a stable as you configure it to be | 12:44 |
xtriz | so is ubuntu-server stable enought to directly use for my work ( i.e. mostly virtualization stuff, running kvm/qemu ) without much fuzz and get my work done without wasting time in fixing things. | 12:45 |
xtriz | ikonia, that's good | 12:45 |
ikonia | xtriz: works out of the box | 12:45 |
xtriz | ikonia, do updates to new kernel and other updates tend to break the system ? ( as you said it's widely used in enterprise so it must not break stuff that easily ) | 12:46 |
ikonia | xtriz: nopr | 12:46 |
ikonia | nope | 12:46 |
TJ- | Hmm, I've always found it doesn't work out of the box... only when I put it *in* the box does it work :) | 12:46 |
xtriz | ikonia, great :) | 12:46 |
Codmadness | TJ- http://pastie.org/9264321 | 12:48 |
TJ- | Codmadness: Aha! so, it's an upstart/configuration issue then | 12:49 |
Codmadness | But still mysql-server isnt installed properly :/ | 12:49 |
TJ- | Codmadness: I thought that part was fixed? | 12:51 |
TJ- | Codmadness: In which case, maybe the update has broken the config files | 12:52 |
TJ- | Codmadness: put the override back in place "echo manual | sudo tee -a /etc/init/mysql.override" ... and then try "apt-get -f install" | 12:52 |
cfhowlett | xtriz suggest you use the LTS (Long Term Support) release - currently 14.04. Tis generally more stable than the interim releases | 12:54 |
xtriz | cfhowlett, ok :) | 12:54 |
=== Den_Beiren is now known as DenBeiren | ||
Codmadness | Ok its doing TJ- | 12:58 |
sander^work | When I install or remove puppermaster and puppetmaster-passenger (from the precise puppetlab repository) I get: invoke-rc.d: initscript puppetmaster, action "start" failed. subprocess installed post-installation script returned error exit status 1 | 13:38 |
Tung | sudo is really really annoying for someone coming from debian | 13:40 |
rbasak | Tung: sudo comes from Debian! It's just the installer default. If you wish, just set a root password and use su. | 13:46 |
* rbasak doesn't like root having a password so got used to it | 13:46 | |
Pici | Or use sudo -i if you need to run a lot of commands in a superuser context. | 13:46 |
rbasak | samiux: if you're using the puppetlab repository, you should probably report a bug with them. We can only fix the Ubuntu repository here. | 13:46 |
Tung | actually i've got a more pressing problem than that at the moment but thanks for the suggestions | 13:47 |
rbasak | Ah yes. I use sudo -i all the time. I see no point in sticking sudo in front of everything if that's what I need to do every command. | 13:47 |
Tung | if i give you guys a dump of dmesg do you reckon you could fish out the problem with me plugging in a WD 1TB external USB 3.0 drive into a usb 3.0 port on an intel DH67GD motherboard? | 13:47 |
Tung | i'll go to pastebin | 13:47 |
rbasak | Tung: sounds more like a kernel issue. | 13:48 |
samiux | rbasak, ???? | 13:49 |
rbasak | Tung: you might find #ubuntu-kernel more useful if so. I hope somebody can help you, but hardware issues like that are pretty hard to debug over IRC IMHO | 13:49 |
Tung | http://pastebin.com/m2kkw2An | 13:51 |
Tung | ok | 13:51 |
Tung | take a look anyway, fdisk -l isnt showing the device | 13:51 |
Tung | this is with a 3 day old updated ubuntu 12.04 | 13:51 |
Tung | server | 13:51 |
InFierno | got usb3 drivers installed? | 13:51 |
rbasak | What does /proc/partitions say? | 13:51 |
Tung | i need usb3 drivers? | 13:52 |
Tung | interesting | 13:52 |
rbasak | Is sdc not the device you're looking for? | 13:52 |
Tung | i'm looking for sdc yes | 13:52 |
rbasak | Is sdc listed in /proc/partitions? | 13:52 |
Tung | but fdisk -l doesn't see it and i cannot mount /dev/sdc1 | 13:52 |
InFierno | Do you have hardware raid on? | 13:52 |
Tung | yes sdc and sdc1 is listed in partitions | 13:53 |
Tung | i have hardware bios raid on yes, raid1 | 13:53 |
rbasak | So then your kernel can see sdc and sdc1, which suggests that it reads the partition table OK. | 13:53 |
Tung | the partition is ntfs but it wont mount as such | 13:53 |
rbasak | OK, so it sounds like you have an NTFS mounting problem. Nothing to do with the hardware. | 13:53 |
Tung | ah there we go | 13:54 |
Tung | it mounted now | 13:54 |
Tung | it froze before for 30 seconds | 13:54 |
Tung | but now it mounted | 13:54 |
Tung | with errors: | 13:54 |
Tung | The disk contains an unclean file system (0, 0). | 13:54 |
Tung | The file system wasn't safely closed on Windows. Fixing. | 13:54 |
Tung | my system might've just been busy, thanks anyway | 13:54 |
Tung | all good i see the partition | 13:55 |
Tung | i'm attempting to tar archive project gutenberg for backup to external hdd, any ideas on a faster method of backing it up to a ntfs filesystem? | 14:03 |
Tung | 33GB in size | 14:04 |
Tung | and then i've got the ubuntu mirror to backup which is roughly 103GB | 14:04 |
Tung | tar -cv ? | 14:06 |
Tung | whoa | 14:06 |
rbasak | tar should be fine if you want a filesystem-level backup. If on rotary disks then it's the number of seeks that will slow it down (if you have a large number of files). | 14:06 |
Tung | lol that just dumped the contents of the gutenberg files to my terminal | 14:06 |
rbasak | Recently I've started using bsdcpio which can do pax format archives and I think may include extended attributes etc. | 14:07 |
rbasak | GNU tar is a little more limited and needs parameters to enable that stuff. | 14:07 |
Tung | i've found that mc (midnight commander) has a decent copy command | 14:11 |
Tung | ill use that for now, not tarring | 14:11 |
Tung | getting a respectable 19.xxMB/sec over usb3 with tons of directories/files in each one | 14:12 |
Tung | 67388 files | 14:13 |
=== Tung is now known as Tungasleep | ||
qhartman | Anyone know of a good way to measure the IO activity of a single disk in a machine? Everything I've looked at can only show aggregate information. | 15:12 |
rberg | iostat | 15:14 |
qhartman | awesome, thx | 15:15 |
qhartman | that's perfect | 15:15 |
rberg | np! | 15:15 |
qhartman | http://stream1.gifsoup.com/view2/3554046/the-more-you-know-o.gif | 15:16 |
rberg | nice | 15:16 |
med_ | caribou, what's the right approach to installing crashdump and arming it configging it in Trusty? I think something changed between Precise and Trusty (if memory serves, I haven't dug in yet today) | 15:38 |
med_ | pmatulis, ^ | 15:38 |
med_ | apt-get install linux-crashdump | 15:38 |
med_ | and then do I need to reserve some kernel memory with a boot kernel cmdline entry and reboot? | 15:38 |
* med_ goes a googling | 15:38 | |
=== Trey is now known as Guest66434 | ||
caribou | med_: yep that's the good way then edit /etc/default/kdump-tools | 15:39 |
med_ | nod | 15:39 |
med_ | ack | 15:39 |
med_ | thanks | 15:39 |
caribou | med_: just googles "crash dump recipe" | 15:39 |
caribou | med_: then you will need to reboot so the crashkernel= boot parameter is taken into account | 15:39 |
med_ | so no changes to that doc since 2013 but probably the change I recall happened early 2013 or late 2012 | 15:40 |
med_ | yep, ack, will do | 15:40 |
med_ | yep, that change doc'd in 13.04 | 15:41 |
* med_ invokes the recipe | 15:41 | |
caribou | med_: the mechanism (i.e. kdump-tools) got introduced then, yes | 15:43 |
oste | what is the proper way to install php 5.6? | 15:52 |
genii | oste: Sort of a trick question. The absolute right way would be to package it from source yourself. But that is outside the scope of actually just getting it installed. I'd probably go with PPA here. | 15:58 |
oste | ppa? | 15:59 |
genii | At https://launchpad.net/~ondrej/+archive/php5-5.6 Disclaimer: Personal Package Archives are supported by their authors and not by Ubuntu | 15:59 |
genii | !ppa | 15:59 |
ubottu | A Personal Package Archive (PPA) can provide alternate software not normally available in the offical Ubuntu repositories - Looking for a PPA? See https://launchpad.net/ubuntu/+ppas - WARNING: PPAs are unsupported third-party packages, and you use them at your own risk. See also !addppa and !ppa-purge | 15:59 |
rbasak | Ondřej is the Debian PHP maintainer, so his PPAs are probably the best ones to use. | 16:00 |
oste | cool | 16:00 |
oste | any idea when 5.6 will make it into ubuntu? | 16:00 |
rbasak | Not before it's released. | 16:01 |
rbasak | So it depends on the PHP project's release schedule, and how that relates to both Ondřej, the Utopic release schedule and my own schedule. | 16:01 |
oste | k thanks | 16:02 |
oste | so what is the procedure for getting this ppa in place? | 16:02 |
oste | deb http://ppa.launchpad.net/ondrej/php5-5.6/ubuntu saucy main ?? | 16:02 |
rbasak | If it isn't available in Debian by 21 August, then it definitely won't make it in Utopic. | 16:02 |
rbasak | Since that's Ubuntu's release schedule deadline. | 16:02 |
oste | i see | 16:02 |
rbasak | https://wiki.ubuntu.com/UtopicUnicorn/ReleaseSchedule | 16:02 |
rbasak | https://launchpad.net/~ondrej/+archive/php5-5.6 has instructions | 16:03 |
=== Lcawte|Away is now known as Lcawte | ||
gema | beisner: ping | 18:21 |
mix | buenas | 19:29 |
mix | hi people! | 19:29 |
mix | i have problem with m | 19:29 |
mix | my ubuntu server | 19:29 |
mix | in zoneminder | 19:29 |
mix | i have a error in mysql | 19:30 |
=== thesheff17_ is now known as thesheff17 | ||
mix | hh | 19:52 |
sarnold | mix: welcome to irc :) irc normally works best if you have something concrete to discuss -- like that error from mysql.. | 19:53 |
=== funkyHat_ is now known as funkyHat | ||
rberg | Hi all i am having a issue where if a block device is unavailable during boot the system ask me to press S to skip M for manual or keep waiting, pressing S or M seems to have no effect at all | 20:34 |
rberg | does anyone know what can be done to proceed? | 20:34 |
pmatulis | rberg: try the 'nofail' option in /etc/fstab | 20:36 |
rberg | great thanks! | 20:36 |
pmatulis | rberg: see 'man mount' | 20:36 |
rberg | I had not noticed that before, thanks | 20:37 |
rberg | any idea on why I was unable to skip? | 20:38 |
pmatulis | rberg: nope | 20:39 |
pmatulis | rberg: it's not the root filesystem is it? | 20:39 |
rberg | nope | 20:39 |
=== RoyK^Work^ is now known as RoyK^Work | ||
rberg | this is very odd that I cant press S M or I.. the keyboard input is accepted when I hit ctrl-alt-del.. | 21:00 |
bekks | rberg: So "s", "m" and "i" are defective. | 21:08 |
rberg | yes | 21:08 |
rberg | is plymouth somehow required for these to work? also this is on 12.04 | 21:09 |
bekks | For the keys? | 21:09 |
rberg | yeah.. seems odd I know.. I ask because its not installed | 21:10 |
rberg | and I really am out of ideas | 21:10 |
bekks | plymouth has nothing to do with defective keys. | 21:10 |
rberg | plymouth came to mind as I know it can display that message over the splash screen | 21:11 |
qman | I have had that problem or something very similar before, I never managed to fix it, I ended up reinstalling | 21:13 |
rberg | this doesn't happen often.. but it made recovering a hardware raid problem harder | 21:15 |
bekks | rberg: does the system ask you, before booting ubuntu? | 21:17 |
rberg | not sure I understand what you mean? before ubuntu bios / post? | 21:18 |
bekks | rberg: Yes - does that happen before ubuntu? | 21:18 |
rberg | the problem was that a device in fstab (sdd) had its underlying raid array fall apart. so yes there was a device problem before ubuntu booted | 21:19 |
rberg | I am booting off of sda | 21:19 |
rberg | I would like to be able to skip or drop to recovery shell if this happens again | 21:21 |
qman | Yeah, what's happening is, when there are no problems with disks everything is cool, but if it has to ask you something like skipping or recovering, it bugs out and won't do anything | 21:23 |
qman | I never figured out why or what the fix is, but I have seen it before | 21:23 |
rberg | ok.. at least I am not the only one! heh. I was able to repair the system with a boot cd, and when sdd was online everything boot up fime | 21:24 |
rberg | but I am left asking why was I unable to use the recovery shell | 21:24 |
qman | Right | 21:26 |
qman | It stumped me for over a year, I also had removed plymouth on the server in question | 21:27 |
qman | I had to upgrade for new hardware support and decided a reinstall was the best way | 21:28 |
pr3d4t0r | Greetings. | 21:30 |
pr3d4t0r | Q. Where can I find out which security advisor corresponds to an automatic update just done on one of my Ubuntu servers? I have the linux-3.xx.xx image version. Thanks in advance. | 21:30 |
pr3d4t0r | Er, advisory | 21:31 |
pr3d4t0r | This came out on the heels of the OpenSSL update from 48 hours ago, so they might be related but I'd like to know the details since this requires a server reboot and I have too many users on board. | 21:31 |
rberg | have a look here www.ubuntu.com/usn/ | 21:32 |
pr3d4t0r | rberg: Checking. | 21:32 |
pr3d4t0r | rberg: Found it - futex issue. Thanks. | 21:33 |
pr3d4t0r | Have a nice wknd, Everyone. | 21:36 |
rberg | qman: fwiw I took another system installed plymouth broke a raid array and was able to hit S to skip! | 21:38 |
rberg | well look at that, mountall depends on plymouth! | 21:46 |
qhartman | anyone know if the proble with excessive fsyncs in apt-get causing problems on btrfs systems is still a thing? | 21:53 |
qhartman | I have some VM's running on a ceph cluster that are showing really poor performance, but mostly when running apt-get. Ceph is using btrfs as the storage filesystem, but I would expect that to be abstracted away so that this wouldn't happen | 21:57 |
qman | qhartman: while I'm not sure about your specific case, fsyncs can't really be abstracted away while maintaining data integrity | 22:06 |
qman | qhartman: normally the workaround is to write those fsyncs to a fast cache or logging device, then bring the slow array up to date asynchronously | 22:07 |
qhartman | qman, right. In this case I was just thinking that there are so many "filesystem" layers between the apt-get process and the actual storage that it might not be relevant | 22:42 |
qhartman | but it seems like it is | 22:42 |
qhartman | all the other tests I'm doing work as expected. It's only apt-get that wrecks everything | 22:43 |
sarnold | qhartman: you could always use libeatmydata | 22:46 |
qhartman | sarnold, heh | 22:51 |
qhartman | interesting | 22:51 |
qhartman | that might be a good validation check | 22:51 |
sarnold | qhartman: fwiw I'd really expect an fsync in a program to really wait until data has been committed to disk, at least one disk, somewhere in the setup. that's what it asks for. hehe. | 22:52 |
qhartman | sarnold, sure. When you're going "program -> OS -> hypervisor -> clustered storage -> OS -> filesystem" I wouldn't be surprised to have something in there lying a little bit. | 23:00 |
=== Tungasleep is now known as Tung | ||
=== beisner is now known as beisner-weekend |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!