kevko | anyone here ? | 05:46 |
---|---|---|
kevko | find a bug ...and i think it's impossible to fix because of containerized env :( | 05:46 |
opendevreview | Bartosz Bezak proposed openstack/kolla-ansible master: Add Pure Storage FlashBlade as Manila backend https://review.opendev.org/c/openstack/kolla-ansible/+/879846 | 07:14 |
opendevreview | Christian Berendt proposed openstack/kolla-ansible master: ceilometer: use template module in "Copying over custom pipeline.yaml file" https://review.opendev.org/c/openstack/kolla-ansible/+/890056 | 07:50 |
opendevreview | Bartosz Bezak proposed openstack/kayobe master: Add STP option for bridge interfaces https://review.opendev.org/c/openstack/kayobe/+/889859 | 08:07 |
opendevreview | Bartosz Bezak proposed openstack/kayobe master: Add STP option for bridge interfaces https://review.opendev.org/c/openstack/kayobe/+/889859 | 08:23 |
mnasiadka | kevko: what bug? | 08:27 |
frickler | apropos bug and containerized, did anyone ever look into virtlogd? it isn't running in our libvirt container currently, but would be really nice if nova/libvirt could use it for console logging | 08:31 |
mnasiadka | frickler: I mentioned that multiple times that it would improve logging - but seems nobody had cycles :) | 08:52 |
frickler | mnasiadka: ah, sorry, I must have ignored that. guess I'll need to find some cycles for it, then | 08:56 |
opendevreview | Michal Nasiadka proposed openstack/kolla-ansible stable/2023.1: Fix the fluentd regexp to collect the logs https://review.opendev.org/c/openstack/kolla-ansible/+/884539 | 09:42 |
opendevreview | Michal Nasiadka proposed openstack/kolla-ansible master: CI: Fail on fluentd log parsing errors https://review.opendev.org/c/openstack/kolla-ansible/+/742399 | 09:43 |
kevko | mnasiadka: we installed nvme package into nova-compute, and when os-brick is initialized ..also nvmeof driver is initialized ... and there is a line which not working in containerized environment .. https://github.com/openstack/os-brick/blob/bfb55b7b58b4a89a690c9c72d7c0ec45b3b66493/os_brick/initiator/connectors/nvmeof.py#L786 | 11:01 |
kevko | mnasiadka: normally in OS it will return uuid disk which is mounted as / ...in containerized environment it returns "overlay" ...so function def _get_host_uuid(self) will fall into except block .... | 11:03 |
kevko | for us it is not problem because we are not using it ...but i would like to fix because it's generating error in logs with start of instance ... | 11:03 |
kevko | i was wondering that we can create wrapper so openstack code will return proper result even if it is in container ... | 11:04 |
kevko | we will just mount /etc/mtab into container to some location ..something as /etc/mtab-docker-host ..and parse from file instead of direct findmnt and then blkid ...what do u think ? | 11:05 |
kevko | OR - we can remove nvme package from images and drop support for it ... | 11:06 |
mnasiadka | well, shouldn't that be fixed in os-brick? | 11:07 |
kevko | mnasiadka: how ? | 11:07 |
kevko | you are in container ... | 11:07 |
kevko | you can't get information about host os ... | 11:07 |
mnasiadka | find more reliable way of getting UUID? :) | 11:07 |
kevko | how ? | 11:07 |
kevko | you are isolated in container :D | 11:07 |
kevko | let me paste it via paste.openstack.org .. | 11:08 |
mnasiadka | dmidecode still works | 11:08 |
kevko | w8 a minute | 11:08 |
mnasiadka | question why the heck they need UUID of the / filesystem :D | 11:08 |
kevko | well, i think they are using it for HOST_ID | 11:08 |
kevko | mnasiadka: this one -> | 11:12 |
kevko | https://paste.openstack.org/show/b2A1pruYM3N7rbl3rws0/ | 11:12 |
frickler | to me that sounds like a bad idea to generate a host id, not sure if that currently works with linux, but there have been systems using a shared / via NFS | 11:25 |
mnasiadka | isn't dmidecode -s system-uuid better? :) | 11:28 |
rockey | dmidecode demands /dev/mem and /sys/firmware/dmi mounted, then runned as privileged container | 11:41 |
rockey | doesn't sound like a nice workaround, but perhaps run it as an environment variable instead? | 11:42 |
opendevreview | Michal Nasiadka proposed openstack/kolla-ansible master: Remove Glance and Neutron TLS proxies from fluentd input config https://review.opendev.org/c/openstack/kolla-ansible/+/890085 | 11:42 |
opendevreview | Michal Nasiadka proposed openstack/kolla-ansible master: Remove Glance and Neutron TLS proxies from fluentd input config https://review.opendev.org/c/openstack/kolla-ansible/+/890085 | 11:45 |
kevko | mnasiadka: well, maybe it is working because if uuid is None suuid is returned https://github.com/openstack/os-brick/blob/bfb55b7b58b4a89a690c9c72d7c0ec45b3b66493/os_brick/initiator/connectors/nvmeof.py#L764-L782 | 11:58 |
mnasiadka | kevko: maybe the error message needs to be fixed ;-) | 12:12 |
opendevreview | Michal Nasiadka proposed openstack/kolla-ansible master: fluentd: fix horizon format regex https://review.opendev.org/c/openstack/kolla-ansible/+/890128 | 13:24 |
opendevreview | Michal Nasiadka proposed openstack/kolla-ansible master: fluentd: fix horizon format regex https://review.opendev.org/c/openstack/kolla-ansible/+/890128 | 13:24 |
opendevreview | Michal Nasiadka proposed openstack/kolla stable/yoga: opensearch: move to yum/apt repos https://review.opendev.org/c/openstack/kolla/+/883716 | 13:37 |
opendevreview | Michal Nasiadka proposed openstack/kolla stable/yoga: opensearch-dashboards: Fix permissions https://review.opendev.org/c/openstack/kolla/+/884375 | 13:37 |
opendevreview | Michal Nasiadka proposed openstack/kolla-ansible master: fluentd: fix horizon format regex https://review.opendev.org/c/openstack/kolla-ansible/+/890128 | 14:21 |
opendevreview | Michal Nasiadka proposed openstack/kolla-ansible master: fluentd: fix horizon format regex https://review.opendev.org/c/openstack/kolla-ansible/+/890128 | 14:30 |
opendevreview | Merged openstack/kolla-ansible stable/2023.1: Fix the fluentd regexp to collect the logs https://review.opendev.org/c/openstack/kolla-ansible/+/884539 | 14:35 |
opendevreview | Simon Dodsley proposed openstack/kolla-ansible master: Update Pure Storage NVMe Cinder driver https://review.opendev.org/c/openstack/kolla-ansible/+/879844 | 14:36 |
opendevreview | Simon Dodsley proposed openstack/kolla-ansible master: Update Pure Storage NVMe Cinder driver https://review.opendev.org/c/openstack/kolla-ansible/+/879844 | 14:42 |
opendevreview | Simon Dodsley proposed openstack/kolla-ansible master: Update Pure Storage NVMe Cinder driver https://review.opendev.org/c/openstack/kolla-ansible/+/879844 | 14:43 |
spatel | mnasiadka what is this option for? - neutron_enable_rolling_upgrade: "yes" | 15:39 |
mnasiadka | for doing rolling upgrade which is the default | 15:39 |
mnasiadka | if you set that to no, we call legacy_upgrade.yml | 15:39 |
mnasiadka | which is not rolling | 15:40 |
spatel | oh okay so better to keep it yes :) | 15:40 |
mnasiadka | frickler, bbezak, mgoddard: maybe we should get rid of those legacy upgrade options? | 15:46 |
mnasiadka | kevko ^^ | 15:47 |
opendevreview | Merged openstack/kolla-ansible stable/2023.1: opensearch: alter path after using rpm/deb packaging https://review.opendev.org/c/openstack/kolla-ansible/+/889806 | 16:29 |
*** mjknoxie is now known as mjk- | 17:22 | |
spatel | Trying to read the kolla image building code and came across this line - PS1="${_KBOLD}(${KOLLA_SERVICE_NAME})${_KRESET}[\u@\h \W]\\$ " | 18:38 |
spatel | what is KOLLA_SERVICE_NAME coming from? | 18:38 |
spatel | oh wait its ENV variable - KOLLA_SERVICE_NAME=keystone-ssh | 18:38 |
spatel | got it | 18:39 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!