[07:11] Good morning [09:27] Is ntpd better then timesyncd? [09:27] mean is it more exact? It is stated here https://www.digitalocean.com/community/tutorials/how-to-set-up-time-synchronization-on-ubuntu-16-04 [09:41] adac: IIRC timesyncd uses only one server to sync from [09:42] adac: and NTP uses by default 4 I think [09:42] that could be (one) part of a difference between them [09:42] adac: although the cases where one is so much better that it matters is IMHO limited to almost scientific cases [09:43] just like chrony being better in this regard was no reason for many people to switch since then [09:43] https://chrony.tuxfamily.org/comparison.html [09:43] I had no numbers on ntpsec, but I'd assume the clenaup of ntp also didn't hurt [09:45] adac: for a longer list of bikeshedding around that read https://news.ycombinator.com/item?id=15324386 [09:46] cpaelzer, I tried now to switch, to ntpd, but I think nothing has changed. I have some problems with a load balancers queue time that gets measured by scoutapp [09:46] the guys thought that the high queue time could be maybe caused by out of sync datetime on my servers [09:46] but I think i have ruled that one out now [09:46] adac: at what level of precision is your issue secs, msecs, µsecs, ... ? [09:47] cpaelzer, milliseconds actually [09:47] yeah, I doubt that was a time sync issue [09:47] it should be about 1.2 ms as on my production server but currently is about 600+ ms on my staging server [09:47] but you are on the same conclusion already, so good luck testing your further theories [09:48] cpaelzer, thanks! And thanks for your hints! [12:12] cpaelzer, ahasenack: I merged nacc's lint fix as it was already reviewed by cpaelzer. It seems to be working in the edge snap now. [12:13] I wonder if I should push the edge snap to beta and/or stable? [12:13] What is everyone using currently? [12:13] we are using edge [12:13] stable is quite behind, I'm not sure what the criteria to push to stable is [12:13] probably that is is stable :) [12:14] Right now the only point of stable, AFAIK, is to give users something to revert to if we break beta. [12:15] (but we'd prefer to users to be using beta or edge so we can get more testing) [12:17] edge as well, ecept for imports obviously [12:19] Yeah all pushed imports via beta please, as that's what the daemon importer runs. [12:19] coreycb: https://bugs.launchpad.net/ubuntu/+source/python-taskflow/+bug/1736394 [12:19] Launchpad bug 1736394 in python-taskflow (Ubuntu) "unit test failure with oslo.serialization 2.21.2" [Undecided,New] [12:19] reproduced in a tox env directly upstream as well [12:28] rbasak: a git question, if I may [12:29] rbasak: I have the landscape-client cloned repo (git ubuntu clone landscape-client) [12:29] I worked on it, have branches, remotes, etc. Now I want to update ubuntu/devel to whatever it's in the archive [12:29] rbasak: I did git fetch pkg --tags --force, and now I'm trying to update the branch [12:29] but it keeps asking me for a commit merge, i.e., it's trying to do a merge commit [12:29] git pull pkg ubuntu/devel <-- that [12:30] even with --ff [12:30] i didn't change ubuntu/devel locally [12:30] at least not knowingly [12:37] If it demands a merge commit, then it's because the current ubuntu/devel (local) is not an ancestor of pkg/ubuntu/devel. [12:37] You could do "gitk ubuntu/devel pkg/ubuntu/devel" to visualise how they differ [12:38] I just tried git pull pkg ubuntu/devel --rebase [12:38] "gitk ubuntu/devel...pkg/ubuntu/devel" may also be helpful - it'll hide unrelated commits. [12:38] and that seemed to have done the right thing [12:39] I now see the changes that are in the archive [12:39] d/changelog is correct [12:39] tec [12:39] with the rich history from the latest upload [12:40] I can investigate further with the help of your reflog to see what was there before if you wish. [12:40] It's possible that the importer did a non-fast-forwarding update to pkg/ubuntu/devel. [12:41] Which shouldn't happen except in the case of an archive admin intervention (a delete in the archive) but I don't think that should have been the case here. [12:41] it's fine [13:06] jamespage: ack, 2.21.2 is causing more problems [13:11] coreycb: I've uploaded a compat patch for taskflow [13:11] might be something similar for glance [13:11] jamespage: ok great [14:02] jamespage: regression tests run successfully against mitaka-proposed. any objections to promoting the following to mitaka-updates? ceph, keystone, libxml2, lxc, python-oslo.concurrency, python-pysaml2, qemu, rabbitmq-server. [14:40] coreycb: no go for it [14:41] jamespage: ok [14:41] lets have a clear through :-) [14:41] coreycb: had a clear through the new queue for pkgs; nothing significant [14:41] jamespage: i'm holding off on a few that haven't landed in xenial-proposed or other reasons [14:42] coreycb: doko has made requests for rmq, ovs and ceph to switch to using py3 for cli tools; I've triaged those as wishlist for 18.04 [14:42] jamespage: ok. maybe we can make some progress during this next milestone. [14:44] coreycb: ok [14:46] jamespage: do you know what the status of vmware-nsx is in mitaka-proposed? the version is ahead of xenial. [14:49] hmm === Jynxie_ is now known as Jynxie [15:56] cpaelzer: actually, here is better [16:10] ahasenack: ok so on nvme? [16:10] cpaelzer: yeah, I'm checking https://bugs.launchpad.net/ubuntu/+source/multipath-tools/+bug/1711749 [16:10] Launchpad bug 1711749 in The Ubuntu-power-systems project "[18.04] multipath-tools: Backport 2 patches to Ubuntu 18.04 (NVMe disks are detected as multipath disks)" [Medium,Triaged] [16:10] cpaelzer: I have a server with nvme [16:10] but I'm not familiar with multipath, and I just learned about nvme-cli (package) [16:12] ahasenack: I didn't come by the nvme-cli package yet either - let me read through that with you [16:12] cpaelzer: I was trying to create multiple namespaces (new concept for me too) in the nvme device to see if I could reproduce the bug [16:20] Office hours officially is started (a bit late), please bring any questions to the ubuntu-server team. [16:26] o/ [16:42] hey slashd [16:42] anything from you this week? [16:43] hey dpb1 everything is under control, was simply showing my presence ;) [16:43] slashd: hehe [16:48] dpb1, Is there a document I must fill ? Like we used to do in the previous meeting format ? [16:48] nope [16:48] dpb1, ack [16:48] we didn't find it provided any value [16:48] dpb1, ok was double-checking in case [16:49] * dpb1 nods [17:59] Greetings. I have a new install of Ubuntu Server. I want to uninstall the OpenJDK but I can't discover the package names. The list scrolls by and off the page. How do I list installed packages one page at a time? [18:01] user | more or | grep "name" [18:04] I'm sorry, that didn't help me. [18:05] Asandari: dpkg --list | more [18:05] what command are you useing ? [18:05] Asandari: Sorry, wrong nickname [18:05] apt list --installed [18:06] andol that worked, thanks [18:06] then use apt list --installed |more [18:07] Thanks guys [21:00] this'll sound like a stupid question but does the server team still hold meetings or have we switched to a more 'office hours' type thing? [21:01] teward: we switched to office hours [21:02] dpb1: got an area where we can put specific hours or no? [21:02] teward: https://community.ubuntu.com/t/irc-meeting-office-hours/1491 -- we also had a mailing list post [21:02] wasn't sure what the decision was, I've been busy :P [21:02] teward: yes! it's the same as before, sec. https://wiki.ubuntu.com/ServerTeam/Meeting [21:02] ah, I see. [21:03] yeah that's smack dab in the middle of things that take my time, for now. That'll hopefully change :) [21:30] hi! Who know how to test server correctly? [21:31] I put two wp naked site on two servers, one on digital ocean and second on my current and my current show server response time 0.4 seconds, on digitalocean less than 0.2. [21:31] What does it mean? [21:32] digitalocean: [21:32] https://developers.google.com/speed/pagespeed/insights/?url=http%3A%2F%2Ftest.american-chat.ru%2F [21:32] Current vps: [21:32] https://developers.google.com/speed/pagespeed/insights/?url=http%3A%2F%2Fkselax.kselax.ru%2F [21:32] it means my current server is sucking? [21:33] I'm not expert, but 0.4 seconds for naked wp site without plugins and load it's big time, with load it will significantly higher [21:45] who know who correct set up permissions for site folder? [21:45] My WP site can't install plugins [21:45] folder belongs the root user, inside folder belongs to test user, and I can't set up any plugins [21:46] folders has right 755, and I did for inside folder right 775 for can reach site using sftp [22:34] Neo1: there's a lot of variables that influence loading performance [22:34] sarnold: server response? [22:35] sarnold: I think current server is sucking [22:35] it without ssd and digitalocean on ssd [22:35] faster [22:37] Neo1: that could be part of it; it could also be that one server is closer to you than the other; another is how quickly your auth DNS servers respond, the chain of DNS responses needed to get to your host and all referenced resources.. [22:38] for example test.american-chat.ru is 33 ms closer to me than kselax.kselax.ru [22:38] sarnold: how long kselax.kselax.ru load for you? [22:38] test this http://cyberforum.ru, it's also in Moscow and works very fast [22:39] heh that takes forever to load, even with noscript blocking many of the scripts [22:39] sarnold: why? [22:40] sarnold: broken site? [22:40] sarnold: this http://kselax.kselax.ru/ [22:40] a few reasons, none of the referenced resources are hosted on sites that my browser has already looked up and has cached addresses for .. part of it felt like the site being built dynamically before being sent [22:41] sarnold: you can use google [22:41] https://developers.google.com/speed/pagespeed/insights/ [22:41] sarnold: better change server IMHO [22:43] Neo1: note that your response speed may change over time. If one datacenter handles mostly timezone-local requests, it might be much slower during business hours than the middle of the night [22:43] 0.4 seconds and low than 0.2 it's big difference, with plugins and content site will work slower. It's without nothing sucks, and what we'll say when with plugins [22:46] sarnold: see in google inspector kselax.ru http://prntscr.com/hjlz1b [22:47] sarnold: and digitalocean http://prntscr.com/hjlz77 [22:47] Neo1: wow, that's a very different interface than I saw in the google inspector :/ [22:47] current host is faster [22:47] sarnold: as you can see current host is faster [22:48] 2.28s vs 1.79, wow [22:48] yes, current faster, but test on google.ru show other results [22:53] seems my current server even better if measure in google [23:04] stokachu: just saw ubuntu announced a partnership with rancher labs to release a kubernetes cloud native platform. will that be based on juju, or something new? [23:06] Neo1: consider two VPSes on the same node in the same datacenter on different IPs. Both are running identical software. One gives me 0.2s, one gives me 0.3s. The two servers are identical, so what's with the extra 0.1s? The extra .1s could be added for any of fifty thousand potential situations with networking between me and the remote server - trusting solely on that statistic for whether the host is faster or not is a poor metric. [23:06] just saying. [23:14] dpkg-reconfigure -fnoninteractive ca-certificates should work with no prompts right?