Centurion_Dan | rrq: o/ | 09:41 |
---|---|---|
rrq | hi | 09:54 |
rrq | Centurion_Dan: ^^^ | 10:00 |
Centurion_Dan | rrq: oops, sorry, stepped away from the computer for a bit to ice our christmas cake from last christmas :-] | 10:28 |
rrq | YEY util-linux:i386 built (!) for unstable | 10:29 |
LeePen | Centurion_Dan: Hi. We still have one (iced) from 3 years ago waiting to be cut! | 10:29 |
Centurion_Dan | LeePen: I hope it was well dowsed with alcohol... | 10:30 |
LeePen | I also apply that at the moment of consumption! | 10:31 |
Centurion_Dan | :-) | 10:31 |
LeePen | rrq: Great. What did you do to fix it? Have the build hosts been updated? | 10:31 |
rrq | I think Dan at least upgraded the kernel... maybe more(?) | 10:32 |
Centurion_Dan | I wish.... my wife is still sans alcohol and gets jealous if I have some while she can't due to breast feeding... | 10:32 |
LeePen | I am sure you are being supportive ;) | 10:33 |
LeePen | rrq: great. I will have a go at building my backlog of packages. | 10:33 |
rrq | nope arm64 still runs out of memory | 10:35 |
rrq | only the i386 is "fixed" | 10:35 |
LeePen | :( | 10:36 |
LeePen | Centurion_Dan: any idea what is going on with arm64 running out of memory? | 10:37 |
Centurion_Dan | just looking at that.... the flags on dpkg-deb are -z6 -Zxz -Sextreme | 10:37 |
Centurion_Dan | perhaps the -Sextreme is pushing lzma over the top.... | 10:38 |
rrq | is jenkins-glue ? | 10:40 |
Centurion_Dan | nah, lack of memory... It's only got 8G and 7G is allocated to other vm | 10:46 |
Centurion_Dan | so there is less then 1G for the host including arm64 builds. | 10:47 |
rrq | no swap on the arm64 ? | 10:48 |
Centurion_Dan | parazyd or maybe alv will need to take a look at that ^^. Might want to prod jaromil too. | 10:49 |
Centurion_Dan | no swap is setup. | 10:49 |
rrq | I suppose you can't give it a swap-on-file | 10:50 |
Centurion_Dan | rrq: trying... but it fails.... btrfs means fallocate fails to actually physically allocate space.... | 11:19 |
Centurion_Dan | even dd isn't good enough... swapfile fails to mount because of holes.... | 11:22 |
rrq | can we run an emulated arm64 elsewhere? | 11:26 |
Centurion_Dan | short answer is .... no, can't make swap.... | 11:27 |
Centurion_Dan | rrq: I could probably run it on my talosII... | 11:27 |
Centurion_Dan | Or I could find a disk and figure out how to actually make my expressobin run devuan. | 11:28 |
Centurion_Dan | I have 3 of those... | 11:28 |
onefang | Or give it twice as much RAM? | 11:28 |
Centurion_Dan | Most of the ram is allocated to other vm's .... I think parazyd needs to take a look at it.... | 11:29 |
onefang | Ah, out of your control then. | 11:29 |
Centurion_Dan | yup... and because of btrfs and kernel < 5.x we can't create an unholy swap file... | 11:30 |
onefang | I have a VM I use for building a large project, I give it 2GB. I know there are larger projects. | 11:30 |
rrq | how about swapping over nfs then ? ( :)) | 11:32 |
Centurion_Dan | to what?? one of the vm's it hosts?? | 11:36 |
rrq | well, as near as possible I suppose | 11:37 |
Centurion_Dan | chances are that will also use btrfs... | 11:45 |
rrq | if there's disk space available, you could create a 2G image file, format as ext2 and mount, then make a swap in that .. ? | 11:47 |
rrq | (btrfs is one of ny white clouds) | 11:49 |
rrq | my | 11:50 |
onefang | I think that was where the "image file has holes" issue came in. | 11:51 |
Centurion_Dan | I intensely dislike btrfs, because it is very systemd like... | 11:51 |
rrq | yes but I'd expect the ext2 layer to deal with that | 11:51 |
rrq | I always make my VM images starting with hole files .. though on top of ext4 of course | 11:53 |
Centurion_Dan | oh well... bed time... | 11:59 |
Centurion_Dan | ... hmm, | 11:59 |
Centurion_Dan | network interface and MTA make boots really slow to start on beowulf. Somebody mentioned a fix for the former... and we really need to get rid of default MTA out of the tasks | 12:01 |
Centurion_Dan | anyway, I'm off to bed. Nighty nite!! | 12:04 |
Joril | Goodnight! | 12:07 |
rrq | LeePen: no cigar on the arm64 .. I'll see if I can make util-linux for beowulf on amd64 and i386 at least ... might need a version bump | 12:22 |
LeePen | rrq: Thanks. maybe parazyd can look at giving it more memory? | 12:24 |
LeePen | I had success using my ~beowulf1 version scheme. | 12:25 |
rrq | yes, would be good. .. util-linux currently has 2.33.1-0.1+devuan1 .. what would the bump be? | 12:26 |
rrq | that one is already on unstable | 12:27 |
LeePen | 2.33.1-0.1+devuan1~beowulf1 | 12:27 |
LeePen | I know it is a bit long, but it means we keep beowulf less than ceres | 12:28 |
rrq | ok I'll use that | 12:28 |
LeePen | Do we need somebody with physical access to the arm64 host? | 12:29 |
rrq | aiui it's a VM, so it needs someone root on the hosting host | 12:30 |
rrq | don't think physical is needed really | 12:31 |
LeePen | OK. So parazyd, jaromil or ...? | 12:32 |
rrq | maybe some more at Dyne. | 12:34 |
LeePen | What's the best way of contacting them? | 12:42 |
rrq | in theory they are lurking on irc, but nowadays probably email is best | 12:48 |
jaromil | devuan-dev mailinglist, but we read | 12:48 |
rrq | :) | 12:48 |
jaromil | I am caught into full day meetings since 3 days now | 12:48 |
jaromil | but rest assured this is not taken care of | 12:48 |
jaromil | ACK on more RAM on the softiron armt64 host | 12:49 |
rrq | LeePen: now gbp wants the "upstream tag" to be upstream/2.33.1-0.1+devuan1 (rather than upstream/2.33.1-0.1) Did you do something special for policykit-1 about that (in gbp.conf) ? | 12:54 |
rrq | (or even upstream/2.33.1) | 12:54 |
jaromil | sorry wording resta ssured I'm reading and following up | 12:57 |
LeePen | rrq: No, I don't think I did anything else. It just worked. | 12:58 |
rrq | biab | 12:58 |
LeePen | rrq: are you sure you did ~beowulfN not +beowulfN? | 12:59 |
LeePen | jaromil: OK. Thanks. | 13:02 |
LeePen | biab | 13:16 |
rrq | LeePen: thanks; a curly - isn't the same as a straight ~ | 14:12 |
LeePen | rrq: :) At least _that_ is working! | 14:47 |
mason | Jjp137: I just looked, and apache2 is still back one build. I was curious if you did any more digging to figure out why, or if there are any other security patches missing. | 19:55 |
Jjp137 | it's not a problem on packages.devuan.org, so maybe it's some weird amprolla thing? I remember this being an issue before | 19:56 |
Jjp137 | yeah it's very similar (if not exactly) to this: http://maemo.cloud-7.de/irclogs/freenode/_devuan/_devuan.2019-05-16.log.html#t2019-05-16T22:48:25 | 20:00 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!