=== jgrimm is now known as jgrimm-afk === kklimonda1 is now known as kklimonda === aluria` is now known as aluria [11:05] hi [11:16] hi [12:47] was there some bug in the mysql client ? [12:56] Can you be more specific? Most software have bugs that are found and fixed all the time. [13:11] could someone take a look at https://bugs.launchpad.net/ubuntu/+source/os-prober/+bug/1384062 and maybe confirm my findings? this would affect not only Ceph, but possibly other services accessing unmounted partitions directly, e.g. databases [13:11] Launchpad bug 1384062 in os-prober (Ubuntu) "os-prober blocks writes to raw partitions" [Undecided,Confirmed] [13:23] jamespage: ^ [13:55] rbasak, erg [13:56] jamespage: does the charm use os-prober or something? I didn't really understand the problem. But it looks like frickler could help with triaging that. [13:57] rbasak, no - but other things do - I think upgrade-grub does [13:57] Ah [14:25] hallyn, the libiscsi MIR completed - are you ok to add that to qemu for next upload? [14:31] jamespage: rbasak: yes, update-grub is the main problem, and it being called implicitly when updating kernel-packages. === jgrimm-afk is now known as jgrimm === cpaelzer_ is now known as cpaelzer [17:01] ddellav, looks like the only things still blocking for python clients atm are python-openstacksdk and python-oslo.config [17:02] coreycb gotcha. There's some blockers for the oslo's i added to the spreadsheet as well [17:02] ddellav, ok [17:20] jamespage, beisner: neutron 7.0.3 and ceilometer 5.0.2 for liberty tested successfully in trusty-liberty-proposed [17:22] coreycb, okies - please markup the bug verfication-done... [17:22] jamespage, all done [17:22] coreycb, ack. thank you. lmk if you need me to do anything on that. [17:22] coreycb, of course :-) [17:42] ddellav, for oslo.messaging we'll need a merge. take a look at the current ubuntu version (pull-lp-source python-pyngus) to see what our current delta is. [17:42] pull-lp-source python-oslo.messaging [17:44] so for that one, we'll need to update the debian repo to the new release and merge with ubuntu [18:19] ddellav, oslo.i18n and oslo.context uploaded [18:31] smb: trivially reproduced with a new xenial vm, [18:31] http://paste.ubuntu.com/15189468/ [18:31] oh, heh, [18:31] ignore [18:32] forgot my cloudinit scripts don't work bc of that (&$E^(*$&A*&A$* rc.S postinst prompt [18:32] hallyn, not re-logged in to gain libvirtd group [18:32] ah [18:33] yeah i normally don't have to [18:35] smb: sigh, no it was probably something about upgrading my laptop that made it happen. dunno. guess we'll stick to reproducing on p8 :) [18:36] hallyn, its a bit odd since I got it installed here without issues (but I should test upgrades) [18:41] Heya guys, a noobish question. Can someone please explain to me how the HDDs split in this server? :/ I know there is a 500GB SSD and a 2TB HDD but that sda/sda2 part lost me.. Another guy said that sdb1 might be a clone of sda1 if I can verify that somehow.. Thanks! http://paste.ubuntu.com/15189379/ [18:50] NwS: sd* means "scsi disk", so scsi disk a, scsi disk b, etc [18:51] NwS: disks are often partitioned; the partitions get the 1, 2, 5, 6, etc.. [18:51] NwS: so sda1 and sda2 are different partitions on one drive; sdb1 and sdb2 are diffferent partitions on the "second" drive, etc [18:52] there's no guarantee that drives will come up in the same order -- sda sdb sdc etc -- so there's the whole pile of /dev/disks/by-id/ and related symlinks [18:52] sarnold, ty for the info mate, shouldn't the sda1+sda2 = sda size? or am I missing something? [18:53] NwS: loads of variables there :) there may be unallocated space on the drive, the formatted filesystems will lose some space, and drive manufacturers use powers of 1000 to advertise their drives but most software uses powers of 1024 when reporting sizes [18:54] sarnold, ah I see. That is a 512GB Samsung 850 PRO if that matters :P [18:55] I didn't setup the server (obviously) but I am trying to figure out how it was "split" and how I can start using the unused space [18:55] NwS: it does feel like there may be ~80 gigs "missing" somewhere; I've never used lsblk before, so I'm not positive on how to read the results though [18:56] can someone tell me what the heck MaaS is? i'm not sure i quite get it. [18:56] !maas | wxl [18:56] wxl: Metal as a Service is a dynamic server provisioning service for scalability. See more about it at https://maas.ubuntu.com. [18:56] NwS: Please pastebin "sudo fdisk -l" [18:57] bekks, ok got some warnings here let me paste it [18:57] lotuspsychje: incidentially, i did read the website about it. i'm still not quite wrapping my head around what it is and how it's different from other solutions. [18:57] wxl: it lets you treat physical machines a bit like a cloud [18:58] sarnold: so you could have a local network of machines and dynamically provision their resources to sort of act as one machine? [18:58] bekks, http://paste.ubuntu.com/15189769/ ty [18:58] wxl: you were doing great right up until "act as one machine" :) hehe [18:59] ooookie dokie [18:59] sarnold: i'll keep scratching my head about it :) [18:59] wxl: you could have most of the machines off most of the time, but when someone needs to submit a huge parallel compute job, turn on all the machines with GPUs, install an OS + application and start them running [19:00] sarnold: ok that kind of makes sense. thanks :) [19:00] wxl: or you could use maas to automatically install openstack, so that you can then use the usual openstack or euca command line tools to launch VMs -- and when new machines are added to the openstack cluster, you can easily tell them if they are going to be storage machines or compute machines or whatever [19:01] if your needs shrink (say, holidays are oven) then you could use maas to autmatically turn off uneeded machines again [19:01] oic! [19:03] wxl: I think most folks use it primarily to get openstack installed; that's a pretty flexible environment, and can easily hand out VMs from 1 vcpu, 1 gig of ram, 1 gig of disk, all the way up to whatever the biggest machines can provide... [19:03] sarnold: how is this different from using juju to get openstack installed? [19:03] wxl: if you were doing scientific computation tasks or something similar that really -needed- to run on bare metalwithout virtualization getting in the way, maas could work [19:04] wxl: heh, that's the thing, juju just treats maas like another cloud instance -- requesting machines with characteristics to deploy software to :) [19:04] oh! [19:04] NOW i think it makes sense. thanks sarnold! [19:04] \o/ :) [19:09] NwS: Since you are using gpt, can you please pastebin gdisk -l [19:09] or parted -l ? [19:13] bekks, got an error with gdisk -l but parted -l worked: http://paste.ubuntu.com/15189925/ [19:15] NwS: Looks good. You're on a HP server, right? Whats the exact issue at that point? [19:15] I -think- you could just use parted to create a new partition on /dev/sda or resize the lvm partition of /dev/sda [19:16] please make sure your backups have been working before starting :) [19:16] No need for new partitions, all space is either in use by partitions or LVM PV. [19:17] bekks: oh? it feels lik ehtere's roughly 65 gigs missing from the ssd [19:17] bekks: it reports 512gb but the end of partition 2 is at 435 [19:17] bekks, Yeah I am on an HP server. I'm trying to figure out if I am currently using the whole available space (because I think I am missing ~200GB on SSD and can't find any space of the HDD) [19:18] sarnold: thats what we have to investigate now. :) [19:18] And in the end move the logs to the HDD since they are eating up all my boot space.. [19:18] The webalizer logs ** [19:19] NwS: Logs arent in the boot space, so they dont eat it up ;) [19:19] Logs are on the sda1 [19:19] NwS: So lets solve that issue first, since thats the real issue. [19:20] Which is 97% full atm (yesterday the server was down for like 5h because of it being full :/) [19:20] NwS: so pastebin "du -sh /" please, so we can narrow down the space issue. [19:22] bekks, got an error: http://paste.ubuntu.com/15190014/ [19:22] ty for the all help btw bekks and sarnold [19:23] NwS: It was my mistake, please use "sudo du -sh /*" [19:24] I'm logged in as su [19:24] got same msg [19:25] But different output. [19:28] Yeah, tried it a few times it gives the exact same msg but with different numbers [19:28] It gives you a total different output, showing all first-level directories in / [19:28] So pastebin "sudo du -sh /*" please. [19:29] the errors in /proc can be ignored [19:29] those are processes that existed when du read the /proc directory but were dead by the time it tried to read those files [19:30] Ah sorry I've missed the * /facepalm [19:30] you'll see that kind of error often, it's a good idea to get familiar with it :) [19:30] hehe [19:31] Here it is http://paste.ubuntu.com/15190101/ [19:32] NwS: And can you pastebin "cat /etc/issue" too, pleasE? [19:32] http://paste.ubuntu.com/15190140/ [19:38] java 8 ist not really available for trusty, or? Is there some backport? [19:41] spm_draget: there is nothing supported. there is this, however, which may suffice for your needs https://launchpad.net/~openjdk-r/+archive/ubuntu/ppa [19:41] sarnold: http://www.ubuntuupdates.org/ppa/webupd8_java [19:41] Thanks, I will take a look! [19:42] bekks: whose project is that? [19:43] sarnold: you mean the webupd8team ppa? [19:43] bekks: yeah [19:43] https://launchpad.net/~webupd8team/+archive/ubuntu/java -- dont know :) [19:44] looks like alin andrei and stephen brandt [19:44] dunno them, dunno if I want them to have root on my computers :) [19:45] sarnold: :D [19:45] sarnold: thats a good question for every ppa. [19:45] bekks: yes, yes it is :) [19:46] was there a bug in the mysql client ? [19:46] bad connections and such ? [19:52] coreycb python-oslo.serialization is ready for review [19:52] everything else is depwait on oslo.config [19:58] bekks, did you see the last paste? just wondering, no hurry or anything [19:59] NwS: And your wbealizer logs are stored where exactly? [20:00] under /var [20:00] Where exactly? [20:01] var/log/virtualmin/ [20:01] Ah :) [20:01] So can you pastebin "ls -lha /var/log/virtualmin/" or will that reveal sensitive data? [20:02] It will reveal a ton of websites :) [20:03] iirc you had 770-ish megs used in /boot -- which might be tight if you've only got one gig for /boot or something; and /var looked like 20-odd gigs of stuff [20:03] you could clean /boot by keeping just two kernels in there: the currently running kernel and perhaps the newest or second-newest, depending upon which kernel you're running now.. [20:03] out of those 20gbs 16 are the webalizer logs :/ [20:03] and apt-get autoclean, apt-get clean, etc may give you a couple of quick gigs [20:04] I did run clean yesterday in order to make some room but that only gave me back like 600MBs [20:05] bummer [20:05] That's why I am looking for the 2TB drive, and then find a way to move the logs over there :P [20:11] NwS: The 2TB drive is perfectly there, as can be seen in your pastebins. [20:11] ddellav, python-oslo.serialization uploaded, thanks [20:12] :) [20:13] bekks, is it possible to find out which folder/s is actually that HDD? :/ [20:13] NwS: Look at "mount". [20:13] As your pastebins show, it isnt mounted. [20:14] Oh, so I can mount some GBs for example? [20:14] And one last Q and I'm off looking it up. When I mount it will I need to restart the server? [20:15] Most likely, you just want to create a PV (if it doesnt exist already) on that disk, add the disk to a volume group, add a volume, create a filesystem on that volume, and a folder for your webalizer logs. [20:15] NwS: Mounting does not need a restart of the server. [20:15] tyvm for all the info and help bekks :) [20:15] no, but if you add a new mount then you'll want to modify /etc/fstab and rebooting afterwards is a good way to make sure you got the change correct :) [20:16] mount -a is a better way to test fstab functionality. [20:16] ty for the info too sarnold :) [20:16] Will do some digging now [20:34] oh, sarnold, I have a question :) [20:34] what is the state of libteam for ubuntu? [20:35] I noticed it was added, more cause it exists in debian [20:35] but was wondering if there would be any config/script for using it, vs the 1 or 2 blog pages I can find, and are not terribly reliable setup processes [20:37] patdk-wk: hmm, when I reviewed it I found a pile of oddities that I didn't like.. I wonder if anyone got around to addressing them? [20:37] no idea, was looking for any new info, since 16.04 is coming around [20:37] but wasn't able to locate anything new last night [20:38] but went and backported the newest version to trusty to see if it helped my test machine at all [20:38] most of the time, my machine works fine [20:38] but sometimes it will boot without any network working :( [20:39] but just figured out what that was, have to reproduce it again, and dig into it [20:39] I think it's the *standard* issue, udev timing [20:39] since it's in universe any 'main' packages can't hard-depend upon it.. anything that might have used it may be patched to disable it, or load it only on demand rather than by default, etc.. [20:40] na, I wouldn't say make it a default [20:40] was more wondering if there might be any plans to update network up scripts to handle it [20:40] like the bonding driver [20:40] not, make use of it by default :) [20:41] though, i think my biggest issue is, I need to make an init script to run and control teamd, vs launching it via a pre/post-up command [20:42] ooof. yeah that sounds like a lot of work.. [20:43] oh well, was annoying me last night :) [20:43] cause I patched that server, and it didn't come back [20:43] if I thought about it more, I would have got on the console, and dianosed it properly [20:43] instead of just kicking it again [20:44] ugh.. sometimes the quick way is also the slow and annoying way :) [20:44] depends on what your solving for :) [20:45] customer impact, or future problems [20:46] true [21:02] ddellav, python-oslo.config uploaded [21:18] Hi! I'm experiencing a weird issue where after installing Ubuntu 14.04.4 Server the default network route is missing from the route table. I've configured some interface bonding and VLAns statically in /etc/network/interfaces. Do you know how can I debug this problem or what could be happening? [21:18] I looked dmesg and syslog but I don't find the problem/I don't have enough knowledge to detect it [21:18] I verified that I defined only one 'gateway' sentence between all the interfaces as I should do [21:18] Sometimes a manual route that I have defined in another of the interface also is missing [21:18] this happens when I reboot the machine, the routes are there intermittently, sometimes yes sometimes no [23:20] coreycb python-oslo.log, python-oslo.middleware, and python-oslo.cache are ready for review & push