| *** Sukhdev has quit IRC | 00:27 | |
| openstackgerrit | Jamie Lennox proposed openstack/diskimage-builder master: Use the dib python to do cleanup https://review.openstack.org/476302 | 01:20 |
|---|---|---|
| *** rfolco has joined #openstack-dib | 01:51 | |
| *** rfolco has quit IRC | 01:51 | |
| *** rfolco has joined #openstack-dib | 01:59 | |
| *** rfolco has quit IRC | 02:00 | |
| *** rfolco has joined #openstack-dib | 02:03 | |
| *** rfolco has quit IRC | 02:03 | |
| *** Sukhdev has joined #openstack-dib | 02:32 | |
| *** Sukhdev has quit IRC | 02:43 | |
| *** lennyb has quit IRC | 03:40 | |
| *** lennyb has joined #openstack-dib | 03:52 | |
| jamielennox | is there a flag to prevent dib from cleaning up it's tmp directories? | 04:06 |
| jamielennox | i feel like something is going wrong with the package installs, but don't know how to check | 04:07 |
| openstackgerrit | Ian Wienand proposed openstack/diskimage-builder master: Additional mount order unit test https://review.openstack.org/476340 | 04:28 |
| *** andreas-f has joined #openstack-dib | 05:01 | |
| johnsom | Ian thanks again for all of the DIB work recently! | 05:21 |
| ianw | jamielennox: not really, i'd usually drop a "bash" call in where i want it to stop | 05:49 |
| jamielennox | ianw: yea, ok - i was doing pdbs but not getting to the problem | 05:50 |
| ianw | johnsom: hmmm, gate is still much more unstable than i want. until i get the image download jobs cached, i think i'll move them to -nv | 05:50 |
| ianw | and we just started hitting shade floating ip issues ... so, urgh | 05:51 |
| jamielennox | ianw: as far as you know should i be able to run dib inside a privileged docker container? | 05:51 |
| ianw | jamielennox: yes, it has been done i believe | 05:51 |
| johnsom | Well, our lives are better at least. | 05:52 |
| jamielennox | ianw: ever seen: Journal size too big for filesystem. on ext4 images? | 05:53 |
| ianw | jamielennox: ummmm ... not that i can remember sorry | 05:53 |
| ianw | possibly, i guess, i'd suspect the file-system being set too small, rather than the journal too big | 05:54 |
| ianw | heh, not that that's much help :) but we do have size calculations in there and not impossible it could be going wrong | 05:54 |
| jamielennox | ianw: right, so this is not something i understand - but i'm guessing that 64M is too much for the size that we're actually building | 05:56 |
| jamielennox | i have nfi how, but it seems to be < 100m | 05:58 |
| ianw | if doing something small, couple of gb range, maybe yes? I doubt we scale it | 05:59 |
| *** chhavi has joined #openstack-dib | 06:06 | |
| *** aparnav has joined #openstack-dib | 06:12 | |
| *** andreas-f has quit IRC | 06:25 | |
| jamielennox | ianw: what's the relationship between /dev/loopX and /dev/mapper/loopYpZ? | 06:55 |
| *** brault has joined #openstack-dib | 06:56 | |
| openstackgerrit | Xinliang Liu proposed openstack/diskimage-builder master: Fix mkfs use wrong label option for vfat https://review.openstack.org/476382 | 07:10 |
| *** andreas-f has joined #openstack-dib | 07:14 | |
| *** hashar has joined #openstack-dib | 08:06 | |
| *** hashar has quit IRC | 09:01 | |
| *** hashar has joined #openstack-dib | 09:04 | |
| *** andreas-f has quit IRC | 09:06 | |
| *** hashar has quit IRC | 09:09 | |
| *** amarao has joined #openstack-dib | 09:34 | |
| *** Administrator__ has joined #openstack-dib | 09:37 | |
| *** Administrator_ has quit IRC | 09:40 | |
| *** aparnav has quit IRC | 10:11 | |
| *** aparnav has joined #openstack-dib | 10:11 | |
| ianw | jamielennox: /dev/loop0 is the raw loop device, /dev/mapper/loop... is setup by kpartx after we have done the partitioning | 10:27 |
| *** chhavi has quit IRC | 11:21 | |
| *** amarao has quit IRC | 11:30 | |
| openstackgerrit | Merged openstack/diskimage-builder master: On suse the python2 dev package is python-devel https://review.openstack.org/475051 | 12:01 |
| *** farosas has joined #openstack-dib | 12:09 | |
| *** aparnav has quit IRC | 12:19 | |
| *** rfolco has joined #openstack-dib | 12:25 | |
| *** andreas-f has joined #openstack-dib | 12:36 | |
| rfolco | hi andreas-f, is the thursday meeting active ? what is the right place to discuss CI permission to vote on dib-images ? | 12:45 |
| *** hwoarang has quit IRC | 12:56 | |
| *** hwoarang has joined #openstack-dib | 12:56 | |
| *** hwoarang_ has joined #openstack-dib | 13:02 | |
| *** hwoarang has quit IRC | 13:04 | |
| *** andreas-f has quit IRC | 13:36 | |
| mordred | ianw: yah - sorry - I'm looking at those - they REALLY confuse me | 15:01 |
| *** hwoarang_ is now known as hwoarang | 15:04 | |
| *** andreas-f has joined #openstack-dib | 17:19 | |
| *** andreas-f has quit IRC | 18:19 | |
| *** andreas-f has joined #openstack-dib | 19:51 | |
| *** rfolco has quit IRC | 20:30 | |
| prometheanfire | simple-init is broken on gentoo (the element), is there a way to pass in pip install options? | 20:50 |
| prometheanfire | ERROR: (Gentoo) Please run pip with the --user option to avoid breaking python-exec | 20:50 |
| prometheanfire | didn't simple-init install glean as a package at one point? | 20:50 |
| *** andreas-f has quit IRC | 20:50 | |
| *** farosas has quit IRC | 21:07 | |
| *** andreas-f has joined #openstack-dib | 21:23 | |
| *** andreas-f has quit IRC | 22:22 | |
| ianw | rfolco: sorry, this is an off week for the meeting | 22:33 |
| ianw | but yes, we are trying to have the meeting regularly | 22:33 |
| ianw | prometheanfire: i can't image that installing glean as --user is what is wanted? that wouldn't seem to work | 22:34 |
| prometheanfire | it doesn't | 22:38 |
| prometheanfire | I'll need to do something else... | 22:38 |
| jamielennox | greghaynes: so it seems like my problem might be related to using du -x in size calculation | 23:09 |
| jamielennox | but i can't tell why different parts of /built are on different filesystems | 23:09 |
| ianw | jamielennox: what do you mean /built? | 23:31 |
| jamielennox | ianw: sorry /built is wrong, i'm refering to this line: https://github.com/openstack/diskimage-builder/blob/master/diskimage_builder/lib/disk-image-create#L330 | 23:34 |
| jamielennox | with the -x i'm getting a tiny image size which is what is causing my journal is too big error | 23:36 |
| jamielennox | after which if i shrink the journal i get disk usage errors because it dib can't copy all the files onto the tiny image | 23:36 |
| jamielennox | do you know if there is a reason for -x there? is dev mounted into that fs? | 23:37 |
| ianw | i don't know exactly why, but that would be the guess ... are you doing some sort of multi-partition thing? | 23:39 |
| jamielennox | ianw: it's in a docker in a kubernetes so god only knows whats under the hood | 23:40 |
| jamielennox | but there's nothing in mount that gives it away | 23:40 |
| ianw | can you just dump du_output so we know what it thinks it is seeing? | 23:42 |
| jamielennox | i had one, i'll have to recreate | 23:46 |
| ianw | maybe if you have a union fs underneath it's du's not accurate? unlikely i'm guessing | 23:49 |
| jamielennox | ah, should have put a mount output at that point as well | 23:50 |
| jamielennox | putting a bash prompt there freezes the whole process | 23:51 |
| jamielennox | grrr, this might be a fs bug: https://patchwork.kernel.org/patch/9706629/ | 23:53 |
| ianw | heh, that sounds very plausible! what an odd case | 23:56 |
| jamielennox | ianw: any idea how to read patchwork to know if that's been merged somewhere? | 23:58 |
| ianw | hey, what do you think this is, gerrit or something | 23:59 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!