| openstackgerrit | Merged openstack/diskimage-builder master: Fix typo in CMDLINE env var for bootloader https://review.openstack.org/448508 | 00:02 |
|---|---|---|
| ianw | mmedvede: do you run ubuntu or debian? | 00:28 |
| mmedvede | ianw: Ubuntu | 00:28 |
| ianw | mmedvede: ok, i'm just seeing if we can do a bit better testing these bootloader elements ... change forthcoming | 00:29 |
| openstackgerrit | Ian Wienand proposed openstack/diskimage-builder master: functests: skip qcow2 generically but add specific test https://review.openstack.org/448837 | 00:57 |
| ianw | mmedvede: ^ ok, actually hoping that fails ppc CI :) | 00:57 |
| ianw | that rolls in an old change | 00:57 |
| ianw | hopefully we both speed up ci and get better coverage all at once | 00:58 |
| mmedvede | ianw: oh, so you only enable it on xenial, that is why you asked? I thought you asked what we use in CI in general | 00:59 |
| mmedvede | ianw: the ppc job currently whitelists only debian | 01:00 |
| openstackgerrit | Merged openstack/diskimage-builder master: Capture output in _exec_sudo https://review.openstack.org/448810 | 01:01 |
| ianw | ahh, ok, we could do it for debian i guess | 01:02 |
| mmedvede | ianw: I can just switch it to xenial | 01:02 |
| mmedvede | that probably is better | 01:02 |
| mmedvede | can't use both at the same time | 01:02 |
| ianw | i'm just thinking because that's what we use in the gate | 01:02 |
| ianw | so has the most eyes on it | 01:02 |
| openstackgerrit | Merged openstack/diskimage-builder master: Add default PPC block-device layout https://review.openstack.org/448811 | 01:03 |
| mmedvede | I am rechecking https://review.openstack.org/#/c/448837 | 01:07 |
| mmedvede | ianw: it is failing on ^ as expected! | 01:11 |
| ianw | mmedvede: that's a good sign :) i don't think testing the qcow conversion path over and over again was helping anything | 01:15 |
| mmedvede | ianw: so it does not find the block-device-ppc64el.yaml | 01:15 |
| mmedvede | I have no opinion on testing qcow conversion in functional tests :) | 01:16 |
| ianw | ohhh, doh, i don't think i meant to add that | 01:19 |
| mmedvede | ianw: right, the ARCH is ppc64el, and file is named as if it is 'ppc' | 01:19 |
| ianw | heh, yeah i didn't mean for that to get into that change | 01:20 |
| mmedvede | you did not mean to add the 10-partitoining change? | 01:20 |
| ianw | nope | 01:23 |
| ianw | accidentally staged it | 01:23 |
| openstackgerrit | Ian Wienand proposed openstack/diskimage-builder master: functests: skip qcow2 generically but add specific test https://review.openstack.org/448837 | 01:24 |
| mmedvede | ianw: do you think it is worth adding distro url autodetection for Ubuntu into dib? I think I can use DIB_UBUNTU_DISTRIBUTION_MIRROR on ppc to be able to test both debian and ubuntu. But maybe it should be done automatically for known arches | 01:28 |
| mmedvede | I guess there is a fine line there on how much should be left as an exercise to the end user | 01:30 |
| ianw | umm, yeah, i feel like it's probably ok ... | 01:34 |
| ianw | but if users want to use specific mirrors etc, they will have to figure that out themselves a bit | 01:34 |
| ianw | but it seems reasonable to choose the ports mirror when arch is set | 01:35 |
| *** pmannidi_ has joined #openstack-dib | 01:40 | |
| *** pmannidi has quit IRC | 01:43 | |
| openstackgerrit | Mikhail S Medvedev proposed openstack/diskimage-builder master: Use correct Ubuntu distro url on non-x86 arches https://review.openstack.org/448848 | 01:57 |
| *** owalsh has quit IRC | 02:20 | |
| *** owalsh has joined #openstack-dib | 02:21 | |
| openstackgerrit | Ian Wienand proposed openstack/diskimage-builder master: functests: skip qcow2 generically but add specific test https://review.openstack.org/448837 | 02:49 |
| *** pmannidi_ has quit IRC | 04:11 | |
| *** aparnav has joined #openstack-dib | 04:23 | |
| *** pmannidi_ has joined #openstack-dib | 04:28 | |
| *** pmannidi_ has quit IRC | 04:43 | |
| *** pmannidi_ has joined #openstack-dib | 05:00 | |
| *** venki has quit IRC | 05:32 | |
| *** andreas-f has joined #openstack-dib | 05:41 | |
| *** venki has joined #openstack-dib | 05:47 | |
| *** venki has quit IRC | 05:54 | |
| *** venki has joined #openstack-dib | 06:07 | |
| *** appu103 has joined #openstack-dib | 06:09 | |
| openstackgerrit | Ian Wienand proposed openstack/diskimage-builder master: [DNM] Try symlinks for grub-install on PPC https://review.openstack.org/448914 | 06:10 |
| openstackgerrit | Ian Wienand proposed openstack/diskimage-builder master: [DNM] Try symlinks for grub-install on PPC https://review.openstack.org/448914 | 06:12 |
| *** aparnav has quit IRC | 06:13 | |
| *** venki has quit IRC | 06:18 | |
| yolanda | hi, so i finally don't need an isolated /boot partition, at least for my security use case. But enabling that possibility is good | 06:36 |
| *** venki has joined #openstack-dib | 06:37 | |
| yolanda | i need partitions for /var, /var/log, /var/log/audit, /tmp | 06:37 |
| *** venki has quit IRC | 06:41 | |
| prometheanfire | yolanda: seperate /usr kthnx :D | 06:43 |
| prometheanfire | lets bring up that debate again | 06:43 |
| openstackgerrit | Andreas Florath proposed openstack/diskimage-builder master: Use stevedore for module config of block device https://review.openstack.org/447090 | 06:44 |
| *** venki has joined #openstack-dib | 06:55 | |
| *** jasondotstar has quit IRC | 07:03 | |
| openstackgerrit | gecong proposed openstack/diskimage-builder master: Typo fix: curent => current https://review.openstack.org/448966 | 07:03 |
| *** jasondotstar has joined #openstack-dib | 07:08 | |
| yolanda | prometheanfire, you need separate /usr? this can be done with all the changes in block device | 07:24 |
| yolanda | prometheanfire, from my side i didn't get /usr partition as a security requirement, you need it for security or for something else? | 07:24 |
| *** venki has quit IRC | 08:08 | |
| *** venki has joined #openstack-dib | 08:13 | |
| *** pmannidi_ has quit IRC | 08:13 | |
| ianw | yolanda: that seems good ... though yeah we kind of have to fiddle things for this ppc issue anyway | 08:13 |
| ianw | wouldn't you need a unencrypted /boot for encrypted root applications? | 08:14 |
| *** Administrator_ has quit IRC | 08:20 | |
| *** Administrator_ has joined #openstack-dib | 08:21 | |
| yolanda | i got requests for an encrypted home so far | 08:39 |
| yolanda | not encrypting the whole system | 08:39 |
| *** makowals_ has quit IRC | 09:00 | |
| *** makowals has joined #openstack-dib | 09:01 | |
| *** openstackgerrit has quit IRC | 09:03 | |
| *** hashar has joined #openstack-dib | 09:03 | |
| *** openstackgerrit has joined #openstack-dib | 09:11 | |
| openstackgerrit | yolanda.robla proposed openstack/diskimage-builder master: Refactor: block-device filesystem creation, mount and fstab https://review.openstack.org/444586 | 09:11 |
| openstackgerrit | yolanda.robla proposed openstack/diskimage-builder master: Add lvm management to diskimage-builder https://review.openstack.org/444403 | 09:13 |
| *** venki has quit IRC | 09:30 | |
| *** venki has joined #openstack-dib | 09:38 | |
| *** venki has quit IRC | 10:11 | |
| *** makowals has quit IRC | 10:38 | |
| *** makowals has joined #openstack-dib | 10:43 | |
| *** appu103 has quit IRC | 11:15 | |
| prometheanfire | yolanda: I was joking, but seperate /boot would be nice | 13:44 |
| prometheanfire | seperate /var /opt /home /tmp (and possibly /usr ) is nice though | 13:45 |
| prometheanfire | mounting with nodev | 13:45 |
| yolanda | well yes, support for any different mountpoints, with different flags, is covered | 13:50 |
| yolanda | /boot is a particular use case, because bootloader element needs to be refactored | 13:50 |
| *** aparnav has joined #openstack-dib | 14:18 | |
| *** aparnav has quit IRC | 14:27 | |
| *** aparnav has joined #openstack-dib | 14:32 | |
| openstackgerrit | Mikhail S Medvedev proposed openstack/diskimage-builder master: Use correct Ubuntu distro url on non-x86 arches https://review.openstack.org/448848 | 14:34 |
| openstackgerrit | yolanda.robla proposed openstack/diskimage-builder master: Refactor: block-device filesystem creation, mount and fstab https://review.openstack.org/444586 | 15:45 |
| *** aparnav has quit IRC | 17:03 | |
| *** hashar has quit IRC | 17:58 | |
| openstackgerrit | Andreas Florath proposed openstack/diskimage-builder master: Refactor: block-device filesystem creation, mount and fstab https://review.openstack.org/444586 | 19:17 |
| yolanda | hi andreas-f , ianw , i started testing on ironic of my whole disk images, but getting that error | 20:13 |
| yolanda | nstanceDeployFailure: Deploy failed for instance 8322306c-a748-4a43-aad7-6000b995fa61. Error: Failed to check the number of primary partitions present on /dev/disk/by-path/ip-192.168.24.9:3260-iscsi-iqn.2008-10.org.openstack:c0c6a5bf-cc82-45df-99c3-30b75b8f57d4-lun-1 for node c0c6a5bf-cc82-45df-99c3-30b75b8f57d4. Error: The device /dev/disk/by-path/ip-192.168.24.9:3260-iscsi-iqn.2008-10.org.openstack:c0c6a5bf-cc82-45df-99c3-30b75b8f57d4-lun-1 does | 20:13 |
| yolanda | not have a valid MBR partition table | 20:13 |
| yolanda | the image is booting ok with virt-install, but fails on ironic | 20:13 |
| yolanda | so need to debug more | 20:13 |
| *** hashar has joined #openstack-dib | 20:30 | |
| yolanda | andreas-f, ianw , so i saw the issue. I cannot define the env vars on my deploy, because it was giving conflicts with nested yamls. Instead, i created an element, with the hope to provide the right env var: https://review.openstack.org/449122 | 20:36 |
| yolanda | however, it doesn't seem to be applied properly, maybe the env var declared too late | 20:36 |
| yolanda | how can i pass the BLOCK_DEVICE_CONFIG properly from an element, is that possible? | 20:37 |
| ianw | yolanda: hmm ... that seems to be one of the old issues of passing things back from elements -> main. env doesn't work | 20:53 |
| yolanda | i saw that vm is just defining a block-device-config.yaml instead | 20:55 |
| yolanda | so trying the same | 20:55 |
| andreas-f | yolanda: please have a look at https://review.openstack.org/#/c/444586/7/diskimage_builder/lib/common-functions line 357ff how the block-device-config.yaml is used in my patch. | 21:02 |
| andreas-f | There is a patch from ian which does something very similar: https://review.openstack.org/#/c/448811/ | 21:02 |
| andreas-f | yolanda: I'm confused about your earlier statement (9:13) that the boot loader is missing an MBR. I thought only a file system image was needed for ironic? | 21:04 |
| openstackgerrit | Andreas Florath proposed openstack/diskimage-builder master: Refactor: block-device filesystem creation, mount and fstab https://review.openstack.org/444586 | 21:07 |
| *** pmannidi has joined #openstack-dib | 22:15 | |
| *** hashar has quit IRC | 23:07 | |
| ianw | mmedvede: hmm, what's the deal with the ppc failures on the ports patch -> http://dal05.objectstorage.softlayer.net/v1/AUTH_3d8e6ecb-f597-448c-8ec2-164e9f710dd6/pkvmci/diskimage-builder/48/448848/2/check/dib-dsvm-functests-devstack-xenial/d665374/ | 23:52 |
| ianw | i'd really like to get the master tree into the state where the only bit failing for ppc is the bootloader changes | 23:52 |
| ianw | ahh, it's passed now. i guess just an old mirror failure | 23:56 |
| ianw | excellent, so xenial-build works there, for now ... we should merge the bootloader test for that and then it will start failing. then we can actually fix it | 23:56 |
| mmedvede | ianw: yeah, I have seen a few of the mirror failures recently | 23:59 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!