| cpaelzer | good morning | 04:51 |
|---|---|---|
| admcleod | i notice the installer doesnt prompt for tz/locale, what is the rationale here? | 05:04 |
| admcleod | cpaelzer: hi! :) | 05:04 |
| cpaelzer | admcleod: wihch installer | 05:28 |
| cpaelzer | subiquity? | 05:28 |
| lordievader | Good morning | 06:09 |
| cpaelzer | hi lordievader | 06:10 |
| cpaelzer | how are you today | 06:10 |
| lordievader | Hey cpaelzer | 06:11 |
| lordievader | Doing good here, how are you? | 06:11 |
| cpaelzer | lordievader: fine as well | 06:16 |
| cpaelzer | a bit rainy, but no floods here at least | 06:16 |
| Assid | hi there | 06:31 |
| Assid | so im considering a setup with a raid 1 using 1 samsung 850 pro and 1 wd green (mlc+tlc) | 06:31 |
| Assid | i was just asking if theres an issue with using different drives altogether | 06:32 |
| Assid | hello? | 06:39 |
| lotuspsychje | !patience | Assid | 06:39 |
| ubottu | Assid: Don't feel ignored and repeat your question quickly; if nobody knows your answer, nobody will answer you. While you wait, try searching https://help.ubuntu.com or https://ubuntuforums.org or https://askubuntu.com/ | 06:39 |
| lotuspsychje | Assid: the ubuntu channels run worldwide on timezones, not everyone is awake at same time | 06:39 |
| Assid | yeah im aware mate.. | 06:40 |
| Assid | lotuspsychje: any idea if this looks like a drive thats failing? https://pastebin.com/CDL05Ark | 06:40 |
| cpaelzer | Assid: yep raw-read and seek should usually not fail non-rarely | 06:41 |
| lotuspsychje | Assid: smart test passed, perhaps also take a look in your syslog for IO issues | 06:41 |
| cpaelzer | unless you had some crazy special test load (unlikely) | 06:41 |
| Assid | cpaelzer: thats what i was thinking.. it shouldnt be that much.. but the other values dont look like it.. | 06:42 |
| Assid | and the smart tests pass fine | 06:42 |
| Assid | i need to do a long test perhaps tonight after work.. | 06:42 |
| cpaelzer | OTOH | 06:43 |
| cpaelzer | it seems that seagate encodes values more special than the tool might know | 06:43 |
| cpaelzer | http://www.users.on.net/~fzabkar/HDD/Seagate_SER_RRER_HEC.html | 06:43 |
| cpaelzer | so take it with a grain of salt | 06:43 |
| cpaelzer | and maybe take a calculus to check the values :-) | 06:44 |
| Assid | uhh.. crazy idiots.. | 06:44 |
| Assid | if thewy wanna do that.. why not provide a patch to smartmontools to be able to READ it back to a standardised format | 06:45 |
| cpaelzer | https://lime-technology.com/forums/topic/31038-solved-seagate-with-huge-seek-error-rate-rma/ | 06:46 |
| cpaelzer | similar discussion | 06:46 |
| cpaelzer | Assid: I assume there is a list of per device quirks already | 06:46 |
| cpaelzer | just nobody yet cared to create one for this device | 06:46 |
| cpaelzer | unfortunately manufacturers still love to hold back specs sometmies | 06:46 |
| Assid | so this device somehow lags slightly everytime i run a command.. while i get its a 4th gen i5, it really shouldnt behave the way it is | 06:47 |
| cpaelzer | ltrace -S and perf are your friends | 06:50 |
| cpaelzer | this could be so many sources | 06:50 |
| Assid | yeah i;ll check that | 06:50 |
| Assid | so what about the raid 1 situation .. using 1 samsung 850 pro and 1 wd green (mlc+tlc) ; would there be an issue with using different drives altogether | 06:51 |
| Assid | man.. im worse than a person with ADD.. | 06:51 |
| cpaelzer | honestly that is a LMGTFY question | 06:51 |
| cpaelzer | TL;DR not recommended but working, has special implications you usually do not want | 06:52 |
| Assid | while most guides say you should use similar drives or similar spec drives.; i can understand the performance of the slower drive would be the final performance of the system | 06:54 |
| lotuspsychje | Assid: see also the ##hardware channel for hardware questions, like combining ssd's in raid | 06:55 |
| lordievader | Assid: Depends a little bit on how it is implemented, but if the write wait until everything is stored on both drives the benefit of an ssd will dissapear. | 07:08 |
| Assid | yes but it will be as quick as the slowest drive in the array.. which would still be fine; since thats an SSD | 07:11 |
| lordievader | Err, the WD green you mention is a regular hard drive right? Or am I misunderstanding the setup you have in mind? | 07:12 |
| lordievader | But yes, it will be as quick as the slowest dive in the array. | 07:12 |
| cpaelzer | lordievader: Assid: I think writes will be as slow as the slowest | 07:19 |
| cpaelzer | reads will be more interesting | 07:19 |
| cpaelzer | as they might end up being served round robin | 07:19 |
| cpaelzer | being slow and fast one (or whatever i t switches) at a time | 07:20 |
| Assid | WD Green SSD .. | 07:20 |
| cpaelzer | Oh there is a green ssd now | 07:20 |
| Assid | i wouldnt put an SSD with an HDD | 07:20 |
| cpaelzer | well then it is probably ok | 07:20 |
| cpaelzer | if characteristics don't differ too much | 07:20 |
| Assid | yes thats why i mentioned MLC+TLC | 07:20 |
| Assid | i want the benefits of RAID without needing to spend too much on MLC drives | 07:20 |
| cpaelzer | in general it is prefered to use different drives anyway, to lower the chance of breaking at the same time | 07:21 |
| cpaelzer | at least not from same production batch | 07:21 |
| Assid | the TLC just helps incase of hw failure on the drive controller of the samsung.. | 07:21 |
| Assid | i also have a VM which has the snapshot of the database every 15 minutes.. incase of total hw failure | 07:31 |
| Assid | once i learn how to use pg replication , i'd probably use that instead of snapshots | 07:31 |
| jamespage | coreycb: I'm going to start a run of dep refereshes for b2 | 09:25 |
| jamespage | coreycb: merging with debian as I go | 09:44 |
| jamespage | coreycb: oslo.config to start with | 09:44 |
| jamespage | coreycb: ok config, utils, log, i18n and oslotest done and uploaded to cosmic | 10:53 |
| jamespage | doing context now | 11:08 |
| rawi | Hello folks. Xenial Server with encrypted root partition. After the kernel update to 4.4.0-128 the server goes into endless boot loop. Booting the old 4.4.0-127 is OK. Somebody here, who experienced it to? | 11:09 |
| jamespage | context done moving to cache | 11:19 |
| coreycb | jamespage: cool, i'm starting on clients now | 12:08 |
| coreycb | jamespage: working on glanceclient and heatclient | 12:14 |
| ahasenack | hi, does anobody know if inotify can catch changes to a symlink's target? | 12:22 |
| ahasenack | like I watch resolv (which is a symlink pointing at resolv1) | 12:23 |
| ahasenack | then ln -sf resolv2 resolv (have resolv -> resolv2) | 12:23 |
| ahasenack | but inotifywait didn't catch that | 12:23 |
| ahasenack | https://pastebin.ubuntu.com/p/TSHx3pR8pW/ | 12:23 |
| ahasenack | hm, there are more operations happening there: https://pastebin.ubuntu.com/p/wKpMVY4dVz/ | 12:27 |
| coreycb | jamespage: looks like glanceclient needs a new keystoneauth so going to do that | 12:46 |
| coreycb | jamespage: keystoneauth and heatclient are done | 13:10 |
| coreycb | jamespage: working on keystoneclient and keystonemiddleware | 13:14 |
| jamespage | coreycb: .cache done moving on | 13:40 |
| coreycb | jamespage: ok. glanceclient is done. | 13:41 |
| jamespage | coreycb: oh I see you already did .concurrency - lemme check for a rev | 13:45 |
| jamespage | gbp:info: package is up to date, nothing to do. | 13:45 |
| jamespage | nice | 13:45 |
| coreycb | jamespage: ok that was probably during b1. keystonemiddleware is done. | 13:47 |
| coreycb | jamespage: keystoneclient is done. moving on to python-neutronclient and python-neutron-lib. | 14:09 |
| jamespage | coreycb: ack | 14:19 |
| jamespage | coreycb: oslo.db underway | 14:44 |
| jamespage | coreycb: ovsdbapp needed an update btw | 14:44 |
| coreycb | jamespage: ack. neutronclient and neutron-lib are done. moving on to novaclient and openstackclient. | 14:46 |
| * lopta downloads Ubuntu Server 16.04.4 for i386 | 15:39 | |
| * compdoc passes the hat for donations so lopta can buy a real computer | 15:45 | |
| lopta | compuguy: This is a test rig that I use for things and stuff. I have another test rig for amd64. | 15:48 |
| coreycb | jamespage: novaclient and openstackclient are done. working on os-brick and os-vif. | 15:55 |
| jamespage | coreycb: ok I've got this far - https://launchpad.net/~ci-train-ppa-service/+archive/ubuntu/3285/+packages | 15:57 |
| jamespage | coreycb: will pickup again tomorrow as have a few calls todo before I EOD | 15:58 |
| coreycb | jamespage: great and thanks! | 15:58 |
| jamespage | coreycb: those are all merge/update for rocky | 15:58 |
| jamespage | coreycb: I took the re-align with Debian where possible decicion paths | 15:58 |
| jamespage | but kept to the upstream release tarball for sources | 15:58 |
| coreycb | jamespage: yep makes sense | 16:10 |
| powersj | nacc: rbasak: added pulling the built snap as an artifact for the nightly and ci jobs example: https://jenkins.ubuntu.com/server/job/git-ubuntu-ci-nightly/28/ | 18:51 |
| === lotuspsychje_ is now known as lotuspsychje | ||
| l4m8d4 | Hello there, I am trying to set up a server in the following way: There are 2 system disks, on each disk should reside a big LUKS container, which contains one mirror of a btrfs RAID11 configuration. Now, I used to install on BIOS systems in the past, where I would then install grub to the beginning of both disks, which worked fine. Now with the new EFI system I have here, grub has to reside in the EFI | 19:23 |
| l4m8d4 | partition right? So how do I set up an ESP on both drives so they are both bootable in case of a hardware failure? | 19:23 |
| l4m8d4 | The server installer only lets me select one boot disk, where it creates the ESP | 19:26 |
| rbasak | powersj: thanks! | 20:16 |
| nacc | powersj: excellent, tyvm | 20:53 |
| nacc | powersj: will that lead to eventual space issues/how will it be pruned? | 20:53 |
| powersj | it should (heh) be fine as the jenkins runs are limited to last 25 runs | 20:54 |
| powersj | that is an extra 5Gb of space and I'm ok with that | 20:55 |
| powersj | 2 jobs * 100mb snap * 25 runs | 20:55 |
| powersj | if it gets bad I can lower the number | 20:55 |
| nacc | powersj: ack, thanks for thinking about it :) | 20:57 |
| l4m8d4 | So, nobody here has a clue regarding multiple ESP? | 21:06 |
| rbasak | nacc: so my plan for now is to upload to edge manually using the artifact after checking the hash built matches origin/master. Sound OK? | 21:25 |
| rbasak | (as a process, every time) | 21:25 |
| rbasak | Though I noticed that particular build failed CI due to the bug we haven't been able to reproduce previously | 21:26 |
| rbasak | If that keeps happening, I could try the edge snapcraft snap instead, which would mean adjusting the CI a little. | 21:27 |
| nacc | rbasak: yeah i think that's reasonable | 21:31 |
| nacc | rbasak: i mean the goal is to fix this bug, right? :) | 21:31 |
| rbasak | Yeah but it's blocked on snapcraft being deterministic and then us being able to give Kyle a reproducer. | 21:39 |
| nacc | right | 21:51 |
| nacc | so for now, edge will just be manually updated by you as we land things in master? | 21:51 |
| nacc | and we'll be leaving beta/stable alone until the bug is fixed? | 21:51 |
| rbasak | I thought I could promote the edge snap directly to beta/stable in the store. | 21:57 |
| rbasak | When it's known good. Rather than using the git branches. | 21:57 |
| nacc | rbasak: ah then yeah you can do that | 22:06 |
| nacc | rbasak: are you going to delete the git branch then? | 22:06 |
| nacc | or would we eventually move back to that? | 22:07 |
| runelind_q | I'm trying to set static DNS servers (ipv6) in an ubuntu1804 container. It seems like changes in 50-cloud-init.yaml don't persist, should I create a new file with just the DNS servers? | 22:14 |
| runelind_q | or what's the best way to go about it? | 22:14 |
| nacc | "Changes to it will not persist across an instance." | 22:18 |
| nacc | runelind_q: that's for getting networking info from the data source | 22:18 |
| nacc | runelind_q: you might want #cloud-init as well | 22:19 |
| nacc | or #netplan | 22:19 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!