March 2021 - Heroes - openSUSE Mailing Lists

Zypper with MirrorCache pilot project
by Andrii Nikitin 23 May '21

23 May '21

(Note: this email was sent weeks ago, but it looks it didn't come trough). Hi all, As part of improvements to download.opensuse.org experience, currently new pilot service is available for using with zypper. https://mirrorcache.opensuse.org Its primary task is to redirect zypper requests to a mirror in client's country, similar to what MirrorBrain project does. You can try to replace in /etc/zypp/repos.d/*.repo http://download.opensuse.org with http://mirrorcache.opensuse.org or https://mirrorcache.opensuse.org If you are in Europe or America, the best experience is expected with these addresses: mirrorcache-eu.opensuse.org<https://mirrorcache.opensuse.org> (Europe) mirrorcache-na.opensuse.org<https://mirrorcache.opensuse.org> (North America) Since it is still a pilot, the service comes without any guarantee or obligations. Reasons to try it may have those who: - live in North America and don't want to be redirected to send cross-continent requests when possible; - want to stick to https connection (mirrorcache will try to find a mirror in the same country, which supports https); - are on ipv4-only or ipv6-only connection (mirrorcache will try to find a mirror which supports ipv4 or ipv6, depending on the request's connection); - redirection from d.o.o leads to a mirror, which actually doesn't have the requested file (it is rare, but can happen); - want to try hosting a mirrorcache instance for own location (e.g. organization, country or continent) and enjoy fast and reliable redirection. (with very modest disk space / HW requirements). Additional reasons may have those who: - want to help with improving d.o.o experience and provide feedback; - are interested in mirror management, for which admin's UI is available, and non-admin UI is planned (so a user can add and manage own mirror without global admin rights); https://mirrorcache.opensuse.org/app/server (login using UI menu and ask me for admin rights if you want to use it or see more advanced controls, like job details). - need some related functionality: it is possible that it can be easily achievable with mirrorcache. Further read about differences from MirrorBrain: https://github.com/andrii-suse/MirrorCache/blob/master/doc/mb_compare.md How caching works: - caching happens on folder level, so e.g., when zypper tries to access file, unknown for mirrorcache - it will be redirected to d.o.o, then a background job will collect info about mirrors. Thereafter further requests to the same folder will be properly redirected to a mirror in the client's country, honoring http/https scheme and ipv4/ipv6 connection used: https://raw.githubusercontent.com/andrii-suse/MirrorCache/master/doc/flow.s… - requests to remodata/repomd.xml are not cached, so zypper always gets the latest version of packages. If you are interested in current source code (most of architecture is stolen from OpenQA): https://github.com/andrii-suse/MirrorCache In case of questions or any feedback do not hesitate to contact me. Regards, -- Andrii Nikitin <andrii.nikitin(a)suse.com> DevOPS Automation and Build Service Engineer SUSE Software Solutions Germany GmbH Maxfeldstr. 5 90409 Nuremberg Germany (HRB 247165, AG München) Managing Director: Felix Imendörffer

5 21

[heroes] has list email suffered some new configuration in recent days?
by Felix Miata 31 Mar '21

31 Mar '21

I haven't changed anything in my email setup, or updated SeaMonkey version, but some email from openSUSE, Debian and Fedora lists lately is causing SeaMonkey to disable the droplist button that enables selection of reply-to-list. I'm forced to reply-all, then deselect the post author's email address. -- Evolution as taught in public schools, like religion, is based on faith, not on science. Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/

3 3

Re: MirrorBrain?
by Bernhard M. Wiedemann 30 Mar '21

30 Mar '21

On 30/03/2021 11.08, Gerald Pfeifer wrote: > On Thu 2021-03-18, medwinz wrote: >> My logic is simple, mirrors cannot catch up with the new release if >> several major updates happen at nearly the same time. > > That, plus the timing of announcements also makes a difference. > > Not everyone here probably follows the factory@ list closely (I, > too, read project@ more regularly), so allow me to share a change > Bernhard has recently made which should also help. > > Thank you, Bernhard! For the record, here is what I found during debugging: After the previous full rebuilds, download.opensuse.org IPv6 address had ~50% packet loss for several hours. My nagios[1] has reported this at 2021-02-16 12:20 UTC (20210212 glibc rebuild published) 2021-03-08 11:25 UTC (20210307 glibc rebuild published) 2021-03-12 18:35 UTC (20210311 KDE update published) 2021-03-13 10:37 UTC (still fallout from 20210311?) 2021-03-13 18:09 UTC (still fallout from 20210311?) On the download.o.o VM, top showed ksoftirqd very busy at 98% CPU and several active rsyncd processes using the CPU. This indicates that syncing to the mirrors was still ongoing and that probably also slowed mirrorbrain mirror scans, so that more traffic was directed towards the downloadcontent.o.o site which happens to be the same machine as mirrorbrain (just another IP). All that syncing, scanning and downloading slowed each other down. Apart from the longer publish delay for large Tumbleweed snapshots, I plan to push the setup of the new widehat.o.o machine, so that it can act as a fallback mirror most of the time and relieve downloadcontent.o.o bandwidth. Ciao Bernhard M. [1] http://nagios.zq1.de/nagios4 => http://nagios.zq1.de/cgi-bin/nagios4/extinfo.cgi?type=1&host=oodownload > ==== forwarded message ==== > From: Bernhard M. Wiedemann <bernhardout(a)lsmod.de> > To: factory(a)lists.opensuse.org > Date: Sun, 28 Mar 2021 15:09:28 +0200 > Subject: factory publish delay [was New Tumbleweed snapshot 20210325] > > > 20210325 was the first large snapshot to trigger my new logic for a > longer publish delay (6h instead of 2h). > > The publisher became active when the repo age was 22203 s > and the repo rsync log had 32114 lines and 2091370 byte. > > This time, download.o.o seems to have had a much better time handling > the load than with previous full rebuilds. Though this seems to have > been only a partial rebuild because there are 62802 entries in factory/repo > > Medium-term we can do more fine-tuning to balance publish delay vs load > of download.o.o > > > Ciao > Bernhard M.

1 0

[heroes] Someone please make an English version of this
by Felix Miata 16 Mar '21

16 Mar '21

https://es.opensuse.org/SDB:NVIDIA_Optimus I'm having no luck figuring out the translator pages with gray on gray text, or interpreting their worthless tiny icons. -- Evolution as taught in public schools, like religion, is based on faith, not on science. Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/

9 39

ditching redmine ?
by Per Jessen 15 Mar '21

15 Mar '21

Sasi Olin wrote: > On Mo, Mär 15, 2021 at 15:33, Lars Vogdt <lars(a)linux-schulserver.de> > wrote: > >> I know that the following is kind of unfair, so please don't take it >> as affront: >> * Are the maintainers of Gitlab and Progress aware of this? > > Progress yes, we only discussed Gitlab in passing, so there I'm not > sure I would like to understand the reasoning behind wanting to ditch redmine in favour of <whatever> ? We only fairly recently spent quite a lot of effort on upgrading redmine. I don't even remember this being discussed ? Maybe I missed that meeting, dunno. Brgds Per -- Per Jessen, Zürich (2.2°C) Member, openSUSE Heroes

3 3

https://lists.opensuse.org/archives/list/heroes@lists.opensuse.org/
by Felix Miata 08 Mar '21

08 Mar '21

1 https://lists.opensuse.org/archives/list/heroes@lists.opensuse.org/ Isn't having the complete mailing list address in the URL an invitation for extra spam? Why not simply https://lists.opensuse.org/archives/list/heroes ??? 2 What would be the reason for "I can get to everything but the forums" in this post?: https://www.linuxquestions.org/questions/showthread.php?p=6228168#post62281… -- Evolution as taught in public schools, like religion, is based on faith, not on science. Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/

5 5

Re: Zypper with MirrorCache pilot project
by Neal Gompa 05 Mar '21

05 Mar '21

On Fri, Mar 5, 2021 at 8:40 AM Andrii Nikitin <andrii.nikitin(a)suse.com> wrote: > > From: Neal Gompa <ngompa13(a)gmail.com> > > MirrorManager is *able* to tolerate divergence, but that leads to a > > much bigger database since you wind up tracking all the files > > independently. That is a noted point you mention in the MirrorCache vs > > MirrorBrain differences. > > Maybe you can give me a link to documentation or source code which clarifies this point, because I don't understand it properly. > Or do you mean that this may be implemented in MirrorManager, but is not actually needed? > You can implement this in MirrorManager if you want to, but in practice it is less important if you're tracking only the necessary files that are a proxy for the state of the mirror (e.g. repodata index files and images like ISOs and disk images). MirrorManager currently implements tracking based on repodata index files for RPM and OSTree repositories[1]. For the openSUSE use-case, that would need to be extended to more types or given a flag to just track all files. There would probably need to be some minor adjustments to the code to treat openSUSE main and OBS content differently. MirrorManager has a heuristic to determine what stuff is present when it checks a mirror to auto enable/disable mirrors for different content subsets[2]. [1]: https://github.com/fedora-infra/mirrormanager2/blob/master/mirrormanager2/l… [2]: https://github.com/fedora-infra/mirrormanager2/blob/master/mirrormanager2/l… > > Fedora's system is configured to track and invalidate based on the > > repomd.xml file, which makes it simple and relatively easy to > > determine whether a mirror is eligible. This is combined with the > > usage of metalinks by default so that the client can select multiple > > mirrors and use them for parallel downloads across different mirrors > > as needed, as well as automatic mirror failover. > > You mentioned repomd.xml - does this assume that mirrors host only repositories? > > > Hmm, but the https://mirrorcache.opensuse.org/download/tumbleweed/repo/oss/repodata/repo… > > URL does not work. > > zypper doesn't use such requests as far as I know, but such feature may/should be added for sure. I probably didn't properly prioritize it yet. > But MirrorCache still lacks some features that MirrorBrain has (ASN matching, country-specific configuaration of mirrors, maybe more) > I was using it for rpm-repos-openSUSE[3] and Mock[4] but reverted it[5][6] because MirrorBrain's metalink handling is totally broken. [3]: https://build.opensuse.org/package/show/Base:System/rpm-repos-openSUSE [4]: https://github.com/rpm-software-management/mock [5]: https://build.opensuse.org/package/rdiff/Base:System/rpm-repos-openSUSE?lin… [6]: https://github.com/rpm-software-management/mock/commit/a7d588c22872d71a2f95… -- 真実はいつも一つ！/ Always, there's only one truth!

1 1

Brief update
by Per Jessen 03 Mar '21

03 Mar '21

I will most likely not be able to take part in our meeting tonight, at the very least I will be late. Here is a brief summary from me: I don't think I have really worked on much openSUSE related in February - only minor maintenance stuff and some tickets. * I tried getting the indexing to work on mailman3, Lars allocated some more memory which meant the process ran a lot longer - until it ran out of disk space. When someone has a moment, /dev/vdb on mailman3 could probably do with twice as much space. progress.o.o #88903 * there is a recent ticket from the KDE admin about failing DKIM checks for reports from bugzilla.suse.com - as well as new one about 'suse.com' users being unsubscribed from factory which could be due to DKIM validations failing. I have not yet looked at it in any detail. progress.o.o #89332, #89227 * to amuse myself, I have added some caching (Expires: header) on forum.o.o, to see if caching the most requested static objects would reduce the traffic. Sofar not much effect, not sure why. -- Per Jessen, Zürich (12.4°C) Member, openSUSE Heroes

5 6

Meeting minutes
by Christian Boltz 03 Mar '21

03 Mar '21

Hello, here are the minutes of today's heroes meeting: status reports: * FreeIPA replica set up * does the new version allow _not_ to add itsself to NS records? Would make our DNS setup (FreeIPA -> pdns -> public nameservers) easier * or manage DNS in pdns (+ PowerDNS-Admin) instead of FreeIPA * darix fixed FreeIPA.i.o.o out of the loop for DNS NS records * pagure packages updated in Factory, code.o.o not updated yet * board requested to publish heroes meeting minutes on heroes MLs * several opensuse.* domains moved to openSUSE DNS servers (should be completed in the next days) * progress on the lists.o.o search index - more RAM helped, now needed a bigger disk * DIKM issues with bugzilla mails and suse.com users on MLs * Ricardo will setup status3.o.o on AWS * we'll make the heroes salt repo (including formulas) public on code.o.o (unless someone objects) Plans: * maybe simplify the DNS setup - see FreeIPA above * switch firewall on scar to nftables (with sets) - hands on meeting next Tuesday 20:00 Nuremberg time * switch forums to discourse (includes mail support, embedding as comment area) - after the vBulletin update Regards, Christian Boltz -- > Ach so: es ist natürlich auch möglich, dass ich irgendwo > einen Fehler übersehen habe ;-) Falsch konfigurierte Christians sind nicht mein Problem. [> Christian Boltz und Ratti in fontlinge-devel]

2 1

nftables hands-on meeting next Tuesday
by Christian Boltz 02 Mar '21

02 Mar '21

Hello, we'll do a hands-on meeting next Tuesday (2021-03-09) at 20:00 CET / 19:00 UTC with the goal to switch the firewall on scar.i.o.o from SUSEfirewall2 to nftables. We'll meet at https://meet.opensuse.org/heroes as usual, and probably use a joint ssh session ;-) As usual, there's a meeting ticket: https://progress.opensuse.org/issues/89368 Regards, Christian Boltz -- Yes, English can be weird. It can be understood through tough thorough thought, though. [https://twitter.com/iowahawkblog/status/594168269759623168]

1 0