[02:33] <DirtyCajun> any hints other than the stupid obvious ones online for speeding up initialization on an mdadm array?
[03:11] <rostam> Hi need help please. I installed a package (apt-get install go-server) then I remove it (apt-get purge go-server). Then after I reinstalled it some of the file from the installed pages are missing, how could I recover please?
[03:20] <dpb1> rostam: what files are missing.
[03:21] <dpb1> DirtyCajun: never looked into it
[03:21] <DirtyCajun> dpb1, im probably just being impatient. Its chugging along pretty nicely i just want to increase it more haha
[03:23] <dpb1> on a lark I googled, and I was shocked to find...
[03:23] <dpb1> dev.raid.speed_limit_max = 200000
[03:23] <dpb1> dev.raid.speed_limit_min = 1000
[03:23] <dpb1> wow
[03:23] <dpb1> never knew that one
[03:23] <dpb1> but those are probably the "stupid obvious" ones. :)
[03:23] <DirtyCajun> yep
[03:23] <rostam> dpb1, files in /var/lib/go-server and /var/go directory.
[03:24] <DirtyCajun> i have them at 200000 and 500000 respectively
[03:24] <dpb1> rostam: so
[03:25] <rostam> dpb1, the files are missing in /var/lib/go-server and /var/go directory does not get created ...
[03:25] <rostam> please help
[03:25] <dpb1> rostam: just testing
[03:26] <dpb1> rostam: what version of ubuntu?  I don't see a go-server
[03:26] <rostam> dph1 I did install and reinsall the package afew times, and all the sudden then this issue happened.
[03:26] <rostam> I am using 16.04 LTS update 3
[03:27] <dpb1> rostam: can you do apt-cache policy go-server and pastebin the results?
[03:27] <dpb1> !pastebin
[03:30] <rostam> dpb1, here it is https://paste.ubuntu.com/p/Md5MB3bJJV/
[03:31] <dpb1> rostam: so, here is my fear
[03:31] <dpb1> rostam: I take it that the data in /var/lib/go-server and /var/go was created by you and you want it back?
[03:32] <rostam> dpb1 yes I did copy some files in those directories and then remove them manually.
[03:33] <dpb1> rostam: ah, ok, so just a matter of the package not *creating* directories and files when it's installed?
[03:35] <rostam> yes I copied some files and directories inside those direcory and then remove them manually.
[03:35] <rostam> Yes it does not created those files inside the directories.
[03:35] <dpb1> ok
[03:35] <rostam> please help I need this working
[03:35] <dpb1> so, idk then, /var is a directory used for "variable" data.  data created by the application
[03:36] <dpb1> you'd be best to check with the application itself, support forums, irc channels, etc.
[03:36] <rostam> so you do not think this is ubuntu specifc issue?
[03:37] <dpb1> no.  the package you are installing is even from a 3rd-party repository from the project itself.
[03:38] <dpb1> rostam: see the 'https://download.gocd.org' lines in that pastebin you sent me?
[03:38] <dpb1> I'd go to gocd.org, they are likely who made the packaging.
[03:38] <rostam> okay thanks so much.
[07:04] <lordievader> Good morning
[09:04] <rbasak> powersj, nacc: some help with git-ubuntu CI please? I don't understand the failures or the current status in Jenkins
[09:18] <demahum> Anyone willing to help me understand ubuntu kernel release schedule?
[09:21] <rbasak> Are you looking for https://wiki.ubuntu.com/Kernel/StableReleaseCadence ?
[09:21] <rbasak> Also try the kernel team in #ubuntu-kernel
[09:24] <demahum> rbasak: ubuntu-kernel sounds very good
[09:24] <demahum> rbasak: and for the cadence, I will take a look at it now
[09:24] <demahum> thanks
[10:17] <jamespage> coreycb: ok telemetry bits all uploaded (ceilometer, aodh, panko) + cinder
[11:10] <jamespage> coreycb: nova done - moving onto glance
[11:39] <jamespage> coreycb: glance done
[11:43] <jamespage> coreycb: doing keystone but...
[11:43] <jamespage> msgpack-python -> msgpack is confusing things
[12:17] <boxrick> Hey, I have a service I have replaced the sysvinit script into a Systemd script. However this is going to cause me problems in the future if that package updates for example
[12:18] <boxrick> Is there any clean way of dealing with this?
[12:24] <rbasak> boxrick: we generally avoid doing that in the lifetime of a stable release.
[12:25] <rbasak> We might do it in a new release, but it'd go in /lib/systemd/system/
[12:25] <rbasak> If you put the same name in /etc/systemd/system/, your file will always override one provided by packaging.
[12:25] <boxrick> Oh ok, previously I simply removed the old service
[12:25] <boxrick> IE actually deleted the file
[12:26] <boxrick> So perhaps just leave it in replace then add the systemd file instead.
[12:28] <rbasak> I'm not sure, but I believe systemd will ignore the init.d script if there is a systemd service unit defined of the same name.
[13:07] <coreycb> jamespage: that's too bad, i was hoping the msgpack transition would just work for debs
[13:09] <coreycb> jamespage: i didn't get as far as I wanted on rc1 yesterday. going to get started shortly and will be poking at qemu for ocata in the background.
[13:39] <coreycb> jamespage: i'm going to focus on the rest of the clients and any other dependencies
[13:43] <jamespage> coreycb: well if I just bump to the newer version it will just work
[14:39] <jamespage> coreycb: I need to test with all rbd's to de-risk - doing that now
[15:32] <jamespage> coreycb: not sure which way to jump on this one
[15:32] <coreycb> jamespage: what's that?
[15:33] <jamespage> coreycb: msgpack-python vs msgpack
[15:34] <coreycb> jamespage: new version not working out?
[15:38] <jamespage> coreycb: kinda - have a plan - testing it now
[15:39] <coreycb> jamespage: ok. i'll be done with the clients once I figure test failures with neutronclient and cinderclient.
[15:40] <jamespage> coreycb: \o/
[15:58] <dpb1> office hours starts.... now
[16:15] <nacc_> rbasak: i belileve i saw earlier that jenkins was out of disk
[16:18] <powersj> should be operational now
[16:52] <powersj> rbasak: re-kicked CI on your merges
[16:53] <rbasak> Thanks!
[17:21] <blackflow> I need to use zfsutils-linux 0.7.5 on 17.10. What would be the recommended procedure, can I use the one in bionic-proposed? Or maybe I should use zfsutils-linux srcdeb, repatch for 0.7.5 and build that on Artful?
[17:31] <nacc> blackflow: why do you "need" to do this?
[17:34] <blackflow> nacc: because 0.6.x has a bug with zfs send | receive from a pool of different version (even though features are not enabled). I cannot send from a pool craeted in FreeBSD, to a pool created in Artful, due to that.
[17:34] <nacc> blackflow: did you file a bug in Ubuntu?
[17:34] <blackflow> the bug is filed somewhere in ZoL github, I've hit it before.
[17:34] <nacc> blackflow: if it's an actual bug, we can SRU the fix in
[17:34] <blackflow> lemme find it
[17:35] <nacc> blackflow: but someone has to file it for that to happen (generally)
[17:35] <blackflow> I believe this was it: https://github.com/zfsonlinux/zfs/issues/5699
[17:36] <blackflow> I know last time I temporarily installed gentoo, compiled newer ZoL and it worked just fine. Now I need that kind of solution on Artful.
[17:37] <nacc> blackflow: https://bugs.launchpad.net/ubuntu/+source/zfs-linux/+bug/1733230 ?
[17:39] <blackflow> nacc: oh yes, that's it. subscribed. thanks.
[17:40] <blackflow> meanwhile... I guess I can't use bionic (built against newer kernel I suppose), so I'll try rebuilding on artful from source debs
[17:41] <nacc> blackflow: https://github.com/zfsonlinux/zfs/pull/6602 and others, are a bit hard for me to parse right now
[17:41] <nacc> blackflow: it would help if in the ubuntu bug you can do an executive summary, and perhaps some testing?
[17:42] <blackflow> executive summary?
[17:42] <nacc> blackflow: what the actual fix is?
[17:42] <nacc> blackflow: it's not 100% clear and I don't have time to figure it out right now :)
[17:43] <blackflow> well the fix I did was on gentoo, I just upgraded from 0.6.x to 0.7.x can't remembre which x it was at the time
[17:44] <blackflow> I have no idea if it's something that could be backported. but anyway, I'm willing to help out, right now I have to migrate some large swaths of data and I'd like to avoid temporary installing gentoo somewhere :)
[17:49] <blackflow> nacc: if I'm not mistaken, this is the pull request fix?  https://github.com/zfsonlinux/zfs/pull/6616
[17:52] <nacc> blackflow: right, but the other is the backport, upstrteam
[18:36] <sarnold> blackflow: please forgive me if you already know this, but you can't mix-and-match the zfs utilities with the kernel module -- there's no versioning in the interface, so your utils need to match your module
[18:54] <blackflow> sarnold: I'm rebuilding entire zfsutils-linux package and will install it and all the deps produced
[18:54] <blackflow> infact, I already did. rebooted, pools imported, but zfs send is segfaulting. gotta figure out why now...
[19:02] <nshire> I can successfully ping my Ubuntu 16.04 LTS VM but all I see is a black screen.
[19:03] <nshire> Prior to the black screen I got the normal ubuntu load screen but then it said "fsckd-calced-msg: press ctrl-c to cancel filesystem checks in progress"
[19:04] <nshire> anyone have diagnostic ideas?
[19:06] <mason> blackflow: Just skimming, but is your send/recv issue feature-related?
[19:07] <nshire> also I have ssh access to the vm still, not sure how
[19:07] <blackflow> mason: nah. sarnold hinted at it. the kernel module is not installed with zfsutils-linux packages. this is not dkms... it's part of linux-image package. so yes, I was building and installing new userland, but the kernel module is still 0.6.x
[19:07] <blackflow> I have to pull in zfs-dkms and spl-dkms from bionic and build those too
[19:07] <mason> ah, ah
[19:07] <mason> Shunning the shiny FTW
[19:08] <blackflow> mason: ordinarily I wouldn't do it, but I have a problem with a bug in zfs 0.6.x where you can't zfs send from a pool of different version, even if newer features are NOT enabled
[19:08] <blackflow> mason: 'zis: https://bugs.launchpad.net/ubuntu/+source/zfs-linux/+bug/1733230
[19:09] <mason> Hrm. I'll have to reproduce it later. I'm mostly rsyncing between FreeBSD and Ubuntu based on the nature of what I move.
[19:09] <mason> blackflow: ty
[19:09] <blackflow> mason: yah but I've got snapshots I want to preserve :)
[19:09] <blackflow> otherwise i'd rsync it and call it a day.
[19:09] <blackflow> mason: and btw this is exactly a problem with moving from FreeBSD-created pool to Ubuntu-created pool.
[19:10] <mason> blackflow: I don't see you explicitly not creating features with your pool creation, and FreeBSD makes features that don't exist in Ubuntu.
[19:10] <mason> If you create the pool with -d do you see the same thing?
[19:10] <mason> In a meeting so I haven't dug in enough - sorry if your bug answers that.
[19:10] <blackflow> mason: I've hit this exact problem few months ago. it's reported on ZoL github. upgrading to 0.7.x fixes it.
[19:11] <mason> Right. Just curious.
[19:11] <blackflow> but last time I used gentoo to build the newer version and move snapshots.
[19:18] <blackflow> huh. Two more yaks appeared waiting to be shaved, lol. Arg, I might just as well upgrade to bionic. :)
[19:18]  * mason makes ch-ch-ch-ch-ch noise.
[19:23] <coreycb> jamespage: that took some digging. finally found out why cinderclient tests were failing. exception=The 'msgpack>=0.4.0' distribution was not found and is required by oslo.serialization.
[19:40] <jamespage> coreycb: I'm working a fix for that one
[19:40] <coreycb> jamespage: ok. wasn't sure if that was dh-python or just need to add to pydist-overrides. but maybe something else.
[19:40] <coreycb> jamespage: i'll hold off
[19:40] <jamespage> coreycb: no I'm trying to get msgpack-python to provide msgpack-python and msgpack
[19:41] <jamespage> coreycb: https://launchpad.net/~ci-train-ppa-service/+archive/ubuntu/3142/+packages
[19:41] <coreycb> jamespage: got it, sound good
[19:45] <nshire> for anyone who may be interested in what I did to fix my blackscreen: I SSHed in and ran the vboxguest additions uninstall script at /opt/vboxguestaddition-[...]/uninstall.sh
[19:45] <sarnold> nshire: wow. I *NEVER* would have come up with that. :)
[19:45] <nshire> I'm so glad you appreciate my attempt to help people who may have run into the same issue
[19:46] <nshire> I wasn't able to bring up a terminal on the vm itself, it wasn't clear if it had booted at all
[19:55] <sarnold> nshire: yeah, who knows, this little tidbit might save someone else the same grief in the future. thanks for reporting back :)
[20:28] <jamespage> coreycb: hmm yeah but its not working