LeePen | I have fixed the bad merge of unstable and the correct libmount-dev is now on pkgmaster. | 08:17 |
---|---|---|
LeePen | I still need to understand properly what caused it. | 08:17 |
LeePen | The problem was multiple transient pkgmaster 404 errors around 2022-09-15 19:04:45. | 08:49 |
LeePen | Anything built since then might need a rebuild. | 08:49 |
LeePen | Fortunately, there is nothing... | 08:50 |
LeePen | I have a fix in testing to prevent this in future. | 08:51 |
bb|hcb | LeePen: Thanks! | 16:19 |
bb|hcb | Now let me see if gpsd will build properly or bark about missing -lsystemd | 16:20 |
LeePen | Looking better | 16:41 |
bb|hcb | BTW. In my sbuild env it builds on chimaera and breaks on unstable... | 16:54 |
bb|hcb | Most probably the libsystemd dev package is not there | 16:54 |
LeePen | We have libsystemd-dev | 16:58 |
LeePen | https://pkginfo.devuan.org/cgi-bin/policy-query.html?c=package&q=libsystemd-dev | 16:58 |
LeePen | Otherwise many things would not be buildable. | 16:59 |
LeePen | I don't use sbuild. Prefer pbuilder | 17:00 |
bb|hcb | Yesterday that was not installable. Let me check if that changed | 17:00 |
LeePen | With the mismerge, lots of ceres would have been broken. | 17:01 |
bb|hcb | Maybe my ceres is also broken somehow: libelogind-compat : Conflicts: libsystemd0 (so libsystemd-dev is not installable) | 17:03 |
LeePen | You can't install libsystemd-dev/libsystemd0 on on a systemd with libelogind-compat, but a build chroot can use | 17:05 |
LeePen | the dummy-logind (non-functional) package to get round the dependencies. | 17:06 |
LeePen | Don't do that on a system where you actually require logind functionality. | 17:06 |
LeePen | So on a production ceres/daedalus system you will need a build chroot. | 17:07 |
bb|hcb | TIL :) | 17:07 |
bb|hcb | Good, all build failures were caused by the mismerge :) | 17:17 |
LeePen | Excellent news. I am testing fixes to prevent it in future. | 17:18 |
bb|hcb | As long as the process uses ssh and the VMs have external IPs that is not fixable - when under heavy scan, sshd rejects connections and becomes unusable/unreliable. Even if there are some checks that catch failed ssh/rsync, it will get the process delayed | 17:22 |
bb|hcb | IMHO, it would be better to allow the involved hosts to share some folder, e.g. via NFS over the internal network... | 17:22 |
LeePen | amprolla doesn't fetch using ssh, it uses http like apt. | 17:26 |
bb|hcb | yes, but pushes via rsync over ssh, IIRC | 17:26 |
LeePen | Yes. I have never seen issues with that step. | 17:28 |
LeePen | It is the dak -> pkgmaster (unmerged) rsync that very occassionally moans. | 17:29 |
bb|hcb | The http(s) one?! | 17:36 |
bb|hcb | Oh, thats rsync over ssh too... And pkgmaster is quite exposed to ssh scan :( | 17:37 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!