| shubjero | Bit of emergency here but does anyone know how I can obtain Neutron 16.3.0 packages for Ubuntu Bionic? 16.3.1 is not an option for us | 00:16 |
|---|---|---|
| oerheks | which series? https://launchpad.net/neutron/+packages | 00:25 |
| shubjero | oerheks: Ussuri | 00:29 |
| shubjero | oerheks: the Ubuntu Cloud Archive time manages backports of OpenStack releases for various distributions of Ubuntu. For example, the last supported release of OpenStack on Ubuntu 18.04 Bionic is OpenStack Ussuri | 00:32 |
| shubjero | After that you must move to 20.04 before you can upgrade to Victoria and beyond | 00:32 |
| shubjero | the Ubuntu Cloud Archive team* | 00:33 |
| oerheks | yeas. i would upgrade, but you might have a reason not too. | 00:35 |
| shubjero | Moving to Ubuntu 20.04 is another huge huge deal. We're just trying to get a stable Ussuri on 18.04 before we start working on a plan beyond that | 00:59 |
| patdk-lap | my 18.04 upgrades to 20.04 have been completely painless | 06:24 |
| maret | what's recommended way to format nvme drives in raid from XFS to ext4 if I am need performance(xfs seems to be unstable for me) ? I am planning to use mkfs.ext4 , I am just not sure what parameters to send if I have 3x 2TB nvme drives, if any | 06:25 |
| patdk-lap | how is the raid setup? | 06:25 |
| maret | what do you mean? | 06:27 |
| patdk-lap | you said you had raid | 06:27 |
| maret | I have raid 0 using mdadm | 06:28 |
| maret | and its xfs and I want to try ext4 while keeping as much of perfomance as possible | 06:29 |
| patdk-lap | you cann't put ext4 ontop of xfs | 06:29 |
| patdk-lap | you would have to reformat it | 06:29 |
| patdk-lap | what chunk size did you configure the raid as? | 06:31 |
| maret | 512 and yeah I plan to reformat | 06:32 |
| maret | I was thinking abouty turning lazy load off because I want to do performance test right away | 06:33 |
| maret | fyi here are details https://paste.debian.net/1201489/ | 06:34 |
| patdk-lap | do just add, -E stride=128,stripe-width=384 to your mkfs.ext4 command | 06:35 |
| patdk-lap | for me I normally do like, mkfs.ext4 -T largefile -E stride=128,stripe-width=384 -L testingfs /dev/mdx | 06:36 |
| patdk-lap | but largefile, small, or whatever depending on how large your avg filesize/4 is going to be | 06:37 |
| maret | btw this is for chia cryptocurrency which does a load of writes if thats make any difference | 06:37 |
| patdk-lap | no, only how large the files are | 06:37 |
| patdk-lap | if it's only making like 4 files | 06:37 |
| patdk-lap | use largefile4 | 06:37 |
| patdk-lap | it gives you more usable space | 06:37 |
| patdk-lap | but less total number of files can be on the disk | 06:38 |
| maret | also I got a tip to mount the drive using noatime,nosuid,nodev,nofail,barrier=0,commit=120,data=writeback,x-gvfs-show which I guess goes to fstab, right now I am mounting the drive with plain mount | 06:38 |
| patdk-lap | yep, always use noatime and you will want to disable the journal, barrier I think does that | 06:38 |
| patdk-lap | the others don't matter so much | 06:38 |
| maret | btw i've checked the files generate on the drive and they are somewhere from 80MB. to 900MB | 06:46 |
| maret | what about lazy loading? | 06:47 |
| maret | budlight btw what about other parameters like noatime and barrier=0 | 07:49 |
| BraveheartBSD | Do hardware raid drives have issues with ext4? An 8tb drive has died with 'superblock' errors | 08:35 |
| enyc | BraveheartBSD: i see have answered you in new place ;p | 09:04 |
| BraveheartBSD | Thank you | 09:08 |
| BraveheartBSD | I'll remake that raid using uuid's and hopefully it will work better | 09:09 |
| enyc | BraveheartBSD: or labels, if appropriate | 09:09 |
| enyc | LABEL=braveheartraid1 or whatever can be used if you e2label the ext4's | 09:09 |
| enyc | but you have to be careful labels are unique | 09:10 |
| BraveheartBSD | Yeah, fstab changing things threw me | 09:10 |
| pikaro_ | Hi! On Ubuntu 18.04.5 LTS, looking at ip addr I see that my IPv6 address is autoconfigured as 2001:db8:dead:beef::/64. IOW, the server is receiving a prefix but not assigning itself an IP. why is that the case, and what is this configured "address" for? it doesn't seem to respond to all addresses under it which was my first guess. is it possible to just have it append its EUI-64? I wouldn't | 10:10 |
| pikaro_ | like to manually configure it as static - there's currently no Netplan file and this is a Docker server so the possibility of breaking something is high. | 10:10 |
| enyc | pikaro_: if its' literally that addres you quoted that is an example address only | 11:56 |
| enyc | pikaro_: ::/64 is just a routing / "all routers" address, any static usable address should normally not have 0:0:0:0 in the last 64bits as this is special address basically | 11:58 |
| pikaro_ | enyc, no, it's the actual prefix provided by the hoster | 11:58 |
| enyc | pikaro_: check bth ip -6 ro and ip -6 addr | 11:58 |
| pikaro_ | and yeah that's what confused me so much... | 11:58 |
| enyc | pikaro_: do you have any ipv6 routes? | 11:59 |
| pikaro_ | that just shows the same address, plus its link-local one | 11:59 |
| enyc | pikaro_: e.g. default via fe80:: smething or other link-local ? | 11:59 |
| pikaro_ | yes, I do, a default route via the provider that's also correct as per docs | 11:59 |
| enyc | pikaro_: i'd imagine smething config-wise is affecting the address allocation | 11:59 |
| enyc | pikaro_: e.g. whats' going on with /proc/sys/net/ipv6/conf/all/accept_ra and other similar togles per-interface (not just "all") and other related controls | 12:00 |
| pikaro_ | yeah the "something" is my problem, I've never set up IPv6 on Ubuntu :/ I'm primarily used to Debian and CentOS | 12:00 |
| pikaro_ | But good to know this is not a normal condition. | 12:00 |
| enyc | depends upon the neworkmanager / systemd net config / /etc/config/network and so on what is being used and what it is doing | 12:01 |
| pikaro_ | no choice but to drill down then | 12:01 |
| enyc | pikaro_: in part I'd be tempted to compare with a new ubuntu 20.04 instead | 12:01 |
| enyc | pikaro_: deb11 LTS to come out soon anyhow | 12:01 |
| databender | Got two focal machines, one is set as a local repo (we limit internet access), and I've got a test client updating off of it, but I keep getting errors during the update about amd64 c-n-f metadata not being found. | 16:25 |
| databender | also seeing minimal packages available, but I've allowed everything but the source links | 16:26 |
| databender | what could be going on here? | 16:26 |
| Guest38 | i have a .bin file, when i run it, it says setup has detected that the required package 'rpm' is not installed, what do i do? | 16:29 |
| tomreyn | databender: sudo apt-get update should tell what it's trying to do and how it's failing to do so. | 16:29 |
| tomreyn | databender: there are debug options for it, see apt.conf(5) | 16:31 |
| tomreyn | Guest38: find a compatible build. | 16:31 |
| databender | yeah, that's when I'm getting the metadata errors. I'll check out the debug options | 16:31 |
| tomreyn | databender: feel free to share the full output (indicating any redactions where they took place), on a pastebin | 16:31 |
| Guest38 | hmmmm... where to get the debug logs | 16:33 |
| Guest38 | this one /var/log/debug ? | 16:33 |
| tomreyn | Guest38: i wasn't talking to you about debug logs. | 16:34 |
| Guest38 | oh thats sad.. :( | 16:34 |
| tomreyn | Guest38: what's the software you are trying to run there? have you checked for a variant or version of it that's compatible to your ubuntu release? | 16:35 |
| Guest38 | acronis. https://sg-cloud.acronis.com/bc/api/ams/links/agents/redirect?language=multi&channel=CURRENT&system=linux&architecture=64&productType=enterprise&login=a63b3b3a-f2ce-4f90-b480-cfad875817b9 | 16:36 |
| tomreyn | this answers question one of two. | 16:39 |
| databender | here's my paste: https://pastebin.com/xAzvYX4u | 16:39 |
| tomreyn | so i guess your local mirror is incomplete | 16:40 |
| databender | it gets some stuff, but I find nothing if I search for apache2 for instance. Here's my mirror.list: https://pastebin.com/Vw2Gvtui | 16:41 |
| databender | Yeah, I figured that, just can't figure out what I'm missing (and I'm a newb here, recently switched from CentOS and kind of figuring it out.) | 16:41 |
| tomreyn | does curl -I http:/10.1.8.30/ubuntu/mirror/us.archive.ubuntu.com/ubuntu/dists/focal-updates/main/cnf/Commands-amd64.xz suggest that the file exists? | 16:43 |
| tomreyn | which software interprets the mirror.list file? | 16:44 |
| databender | Nope, throws a 404; shouldn't this be included in the focal-updates main? | 16:44 |
| databender | apt-mirror, right? | 16:44 |
| tomreyn | i see | 16:45 |
| tomreyn | http://us.archive.ubuntu.com/ubuntu/dists/focal-updates/main/cnf/Commands-amd64.xz does exist | 16:45 |
| tomreyn | and you're saying that | 16:45 |
| tomreyn | http://10.1.8.30/ubuntu/mirror/us.archive.ubuntu.com/ubuntu/dists/focal-updates/main/cnf/Commands-amd64.xz does not exist? | 16:45 |
| tomreyn | if so, then your mirror seems to not mirror things properly, yet | 16:46 |
| tomreyn | on the other hand, maybe you don't need the c-n-f metadata at all | 16:46 |
| databender | Yeah, I just figured that there was some link between the 404 errors and the lack of available packages when I use apt-cache | 16:47 |
| databender | I just don't know what else I can mirror other than backports, proposed and source; figured (most) everything would be available in the main focal links | 16:48 |
| databender | I'll try pulling proposed and backports and see if it clears up. | 16:49 |
| tomreyn | bug 1825755 | 16:49 |
| ubottu | Bug 1825755 in apt-mirror (Ubuntu Groovy) "apt-mirror in all versions (including disco and last from github) doesn't mirror Commands-* files" [High, Triaged] https://launchpad.net/bugs/1825755 | 16:49 |
| databender | well, that explains that part; just need to figure out where they keep all the software I'm missing. | 16:50 |
| tomreyn | my experience with the various mirroring scripts has always been that they're always broken one or the other way. but then i last tried this some 10 or so years ago. | 16:51 |
| tomreyn | try asking in #ubuntu-mirrors (after reading the channel /topic) as tho which approach they recommend for a private mirror. | 16:52 |
| databender | Right on, I appreciate the help! | 16:52 |
| tomreyn | you're welcome | 16:52 |
| sigv | tomreyn: we use debmirror on Focal. | 18:47 |
| sigv | (not a part of #ubuntu-mirrors so I am not sure of the state of the discussion there, but wanted to share.) | 18:48 |
| sigv | need some command-line configuration flags - what distros you want to pull and some more things are helpful, but generally 'works'. :) | 18:48 |
| tomreyn | sigv: so you're not running into bug 1825755, as databender does? | 20:27 |
| ubottu | Bug 1825755 in apt-mirror (Ubuntu Groovy) "apt-mirror in all versions (including disco and last from github) doesn't mirror Commands-* files" [High, Triaged] https://launchpad.net/bugs/1825755 | 20:27 |
| tomreyn | oh wait, debmirror isn't apt-mirror | 20:28 |
| sigv | tomreyn: that bug is for `apt-mirror` -- so we are not running into it, by using `debmirror`. | 20:28 |
| tomreyn | yeah sorry | 20:28 |
| sigv | tomreyn: the respective cnf bug for debmirror was LP: #1821251 | 20:29 |
| ubottu | Launchpad bug 1821251 in apt-mirror (Ubuntu) "please add cnf support to debmirror" [Undecided, Confirmed] https://launchpad.net/bugs/1821251 | 20:29 |
| tomreyn | right, i had spotted that one, too | 20:29 |
| tomreyn | you say "was", but this looks unfixed? | 20:29 |
| tomreyn | ah, no it is fixed for debmirror | 20:30 |
| sigv | yeah, the same issue was filed as the same bug number on multiple packages. somewhat misleading here, as the underlying code is different, but hey. | 20:30 |
| sigv | if i link it directly, maybe it picks up debmirror status? https://bugs.launchpad.net/ubuntu/+source/debmirror/+bug/1821251 | 20:31 |
| ubottu | Launchpad bug 1821251 in apt-mirror (Ubuntu) "please add cnf support to debmirror" [Undecided, Confirmed] | 20:31 |
| sigv | not ubottu does not. | 20:31 |
| tomreyn | neither ubottu nor your web browsers' title bar | 20:31 |
| sigv | $ debmirror -a amd64,i386 -d focal,focal-security,focal-updates,focal-backports -s main,universe,restricted,multiverse --nosource --passive --method=http -h archive.ubuntu.com -r ubuntu --keyring=/usr/share/keyrings/ubuntu-archive-keyring.gpg /opt/debmirror/ubuntu | 20:32 |
| sigv | that is what we run on Focal as of now, it seems. I think `-r ubuntu` is not needed, others might be no-op'ish too. | 20:32 |
| sigv | but it works so I haven't bothered touching it too much. :) | 20:32 |
| tomreyn | i'll happily tell databender if i ever see them againaround here AND (less likely) remember about it. | 20:33 |
| tomreyn | or you could memoserv them | 20:33 |
| sigv | oh my. I just realized I pinged the wrong person. -_- | 20:33 |
| tomreyn | it was of interest to me as well | 20:33 |
| tomreyn | thanks | 20:33 |
| sigv | tbh, i would suspect.. most "real" mirrors (a la audience of #debian-mirrors) just FTP the whole thing | 20:34 |
| tomreyn | sigv: https://wiki.ubuntu.com/Mirrors/Scripts | 20:36 |
| sigv | rsync protocol. touché. | 20:37 |
| sigv | learned something myself! | 20:37 |
| tomreyn | only a partial improvement over ftp | 20:37 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!