Sunday, 2019-02-24

*** hwoarang has quit IRC01:11
*** hwoarang has joined #openstack-dib01:13
*** hwoarang has quit IRC02:51
*** hwoarang has joined #openstack-dib02:55
*** hwoarang has quit IRC07:33
*** hwoarang has joined #openstack-dib07:35
*** lennyb has quit IRC08:02
*** lennyb has joined #openstack-dib08:16
fungii wonder if we should consider switching our debian/ubuntu minimal images to build with https://gitlab.mister-muffin.de/josch/mmdebstrap/13:39
*** hwoarang has quit IRC13:41
*** hwoarang has joined #openstack-dib13:42
*** hwoarang has quit IRC15:10
*** hwoarang has joined #openstack-dib15:12
*** brault has quit IRC19:44
ianwcgoncalves: yeah, i feel like it's a bit weird to merge as is because it will disappear?21:15
cgoncalvesianw, sure, I understand. just wanted to know what your plan is :)21:16
cgoncalveswe have validated in Octavia side that we can get an amphora built based on RHEL 8 and passes tempest tests21:17
ianwi'm open to discussion :)  would a branch work?21:17
ianwthere's also that centos8 will have quite a few touch points, so it might be better to get that done on the same branch and then merge when it's working21:18
ianwrather than X point-releases as we fix up little things21:18
cgoncalvesdo you know if there are plans to have RHEL 8 released before feature freeze? I guess not21:18
cgoncalvesI would like to have https://review.openstack.org/#/c/638581/ merged but it depends on your open patch21:19
ianwheh, even if i did know, i don't think i could say out loud :)  but i don't know :)21:19
cgoncalvesif we miss feature freeze, I'll need to carry it downstream which is less than ideal21:19
cgoncalvesno-one is listening ;)21:20
* johnsom leans in21:21
ianwi'll try having a play with glean and see if we can get a -src job working with rhel8.  in that case we could have an infra node21:24
cgoncalvesFWIW, RHEL8-based amphora installs and enables deprecated network-scripts21:25
cgoncalveswe haven't had the chance to support Network Manager in Stein21:26
ianwyeah, you guys are a bit more complex; for glean we use the nm scripts plugin stuff which works with minimal fuss21:27
ianwhrm, either this is broken, or image builds are broken : http://grafana.openstack.org/d/NcFtwKriz/nodepool-dib-status?orgId=121:30
ianw /usr/local/lib/python3.5/dist-packages/diskimage_builder/lib/disk-image-create: line 140: cannot create temp file for here-document: No space left on device ...21:31
ianw786G of images ... which actually seems about right :/  it adds up having vhd+raw+qcow * 12 images * (1 active + 1 old)21:36
ianw /dev/mapper/main-nodepool 1008G 1008G     0 100% /opt/dib_cache/gentoo/portage/packages21:48
ianwboth nb01 & nb02 had mounts like this21:48
clarkbya that looks worse than it is due to lvm aiui21:48
clarkbits part of the main partition and since it is full looks full21:49
ianwyeah, the common thread was it being gentoo mounts21:49
clarkblast time we filled up it was due to leaked images21:49
clarkbwe had images on disk the clouds didnt know about21:49
ianwi'm clearing out dib_tmp on nb01 and investigating a bit on nb0221:49
ianwit does look like a few leaked, particulary debian-stretch : http://paste.openstack.org/show/745951/21:52
ianw images=$(curl -s 'nl01.openstack.org/dib-image-list' | grep -- '-0000' | awk '{print $2}' | grep -o '[^-]*$' | paste -sd "|" -)21:59
ianw ls -c1 *.qcow2 /opt/nodepool_dib | egrep -v "(images)"22:00
ianwshould be a list of stuff to delete that has leaked22:00
ianw ls -c1 /opt/nodepool_dib/ | egrep -v "(images)"22:01
ianweven22:01
*** hwoarang has quit IRC23:09
*** hwoarang has joined #openstack-dib23:13
*** hwoarang has quit IRC23:55
*** hwoarang has joined #openstack-dib23:57

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!