| opendevreview | Steve Baker proposed openstack/ironic master: Capture [pxe]loader_file_paths for distros https://review.opendev.org/c/openstack/ironic/+/815392 | 00:00 |
|---|---|---|
| opendevreview | Steve Baker proposed openstack/ironic master: WIP use privsep to copy for loader_file_paths https://review.opendev.org/c/openstack/ironic/+/821363 | 00:00 |
| opendevreview | Steve Baker proposed openstack/bifrost master: Support PXE network boot with grub https://review.opendev.org/c/openstack/bifrost/+/807220 | 00:25 |
| *** pmannidi|Gone is now known as pmannidi\AFK | 00:34 | |
| *** pmannidi\AFK is now known as pmannidi|AFK | 00:34 | |
| opendevreview | Steve Baker proposed openstack/ironic master: WIP use privsep to copy for loader_file_paths https://review.opendev.org/c/openstack/ironic/+/821363 | 03:40 |
| *** amoralej|off is now known as amoralej | 07:01 | |
| *** akahat is now known as akahat|ruck | 07:25 | |
| arne_wiebalck | Good morning, Ironic! | 08:07 |
| holtgrewe^gone | good morning | 08:09 |
| *** holtgrewe^gone is now known as holtgrewe | 08:09 | |
| jingvar | morning | 08:12 |
| rpittau | good morning ironic! o/ | 08:18 |
| janders | good morning arne_wiebalck holtgrewe jingvar rpittau and Ironic o/ | 08:36 |
| rpittau | hey janders :) | 08:36 |
| arne_wiebalck | hey janders holtgrewe jingvar rpittau o/ | 08:36 |
| rpittau | hey arne_wiebalck :) | 08:36 |
| holtgrewe | arne_wiebalck: \o -- I think I'm 99% there regarding the UEFI image. However it turns out that the root file system is mounted as readonly . /etc/fstab says `LABEL=root_fs / xfs defaults 0 1` and /pro/cmdline is `BOOT_IMAGE=/boot/vmlinuz-3.10.0-1160.49.1.el7.x86_64 root=LABEL=root_fs ro console=tty0 crashkernel=auto net.ifnames=0 console=ttyS0 console=tty0 console=ttyS0,115200 no_timer_check | 08:50 |
| holtgrewe | nofb nomodeset gfxpayload=text rd.auto=1` dmesg says `XFS (md126p3): ro->rw transition prohibited on unknown (0x4) ro-compat filesystem`. | 08:50 |
| holtgrewe | Heh, maybe it's caused by ipa using CentOS 8 and me trying to use CentOS 7... and v8 is using a too recent features. | 08:55 |
| opendevreview | Aija Jauntēva proposed openstack/ironic bugfix/19.0: Fix redfish update_firmware for newer Sushy https://review.opendev.org/c/openstack/ironic/+/822241 | 09:10 |
| arne_wiebalck | holtgrewe: o/ your commandline does not have/need the raid auto-assemble? | 09:10 |
| opendevreview | Aija Jauntēva proposed openstack/ironic stable/xena: Fix redfish update_firmware for newer Sushy https://review.opendev.org/c/openstack/ironic/+/822242 | 09:10 |
| arne_wiebalck | holtgrewe: that should only be a problem the other way round: a c7 IPA installing a cs8 image | 09:11 |
| holtgrewe | arne_wiebalck: actually it's my local Ubuntu 20 machine buidling the image. | 09:13 |
| holtgrewe | arne_wiebalck: OK, can confirm that I got it working. If I want to create some tutorial-style documentation for what it takes to build an image, where is the right place to put it? This repository? https://opendev.org/openstack/diskimage-builder ? | 09:37 |
| arne_wiebalck | holtgrewe: I think it should go to the Ironic docs | 09:39 |
| arne_wiebalck | holtgrewe: https://opendev.org/openstack/ironic/src/branch/master/doc/source | 09:39 |
| arne_wiebalck | holtgrewe: we do not have a place yet for "best pratices / tips and tricks", but it is on my to-do list in fact | 09:40 |
| arne_wiebalck | holtgrewe: I think your recipe will help others a lot | 09:41 |
| arne_wiebalck | holtgrewe: it should probably be on the same level as "Troubleshooting FAQ" | 09:50 |
| arne_wiebalck | holtgrewe: I can try to put sth then you can add (if that sounds ok?). Not sure I manage before the break, though. | 09:51 |
| holtgrewe | arne_wiebalck: I have some experience with sphinx from other projects. No worries, I need to figure out 1-2 other things before. | 09:53 |
| holtgrewe | cloud-init does not work yet and no "centos" user is created by default... | 09:54 |
| * dtantsur rubs his eyes | 10:07 | |
| holtgrewe | Ah, centos is created by cloud-init... | 10:09 |
| dtantsur | holtgrewe: re docs: I think updating https://docs.openstack.org/ironic/latest/user/creating-images.html is a good idea | 10:37 |
| holtgrewe | dtantsur: OK, will look into this. It will probably take me until Thursday to get to it, though. | 10:38 |
| dtantsur | holtgrewe: it's a holiday season, don't rush :) | 10:39 |
| dmellado | dtantsur: o/ I come to bother again, sorry xD | 10:40 |
| dtantsur | hey hey dmellado | 10:40 |
| dmellado | so, trying bifrost w/centos8 it afils when building the image | 10:40 |
| dmellado | so not sure if this is a known issue | 10:40 |
| dmellado | https://paste.openstack.org/show/811765/ | 10:40 |
| dmellado | I tried setting export DIB_INSTALLTYPE_pip_and_virtualenv=package | 10:40 |
| dmellado | didn't really make a difference | 10:41 |
| dtantsur | dmellado: do you see where the actual error is? I cannot find it | 10:41 |
| rpittau | I was going to ask the same | 10:41 |
| dmellado | me neither, that's why I thought it may be something obvious that I didn't understand | 10:42 |
| dmellado | xD | 10:42 |
| dtantsur | dmellado: could it be that it just failed downloading the base image? | 10:42 |
| dmellado | https://paste.openstack.org/show/811766/ | 10:42 |
| dmellado | that's the full start | 10:42 |
| dmellado | stderr seems empty | 10:42 |
| dtantsur | dmellado: try running the same command manually? | 10:43 |
| dmellado | I'll try to run the command manually and see | 10:43 |
| dmellado | yeahj | 10:43 |
| dmellado | a-ha | 10:44 |
| dmellado | network | 10:44 |
| dmellado | it's always dns | 10:44 |
| dmellado | 2021-12-20 10:43:49.941 | Getting /home/dmellado/.cache/image-create/CentOS-8-GenericCloud-8.4.2105-20210603.0.x86_64.qcow2.tgz.lock: Mon Dec 20 04:43:49 CST 2021 | 10:44 |
| dmellado | 2021-12-20 10:43:49.942 | Fetching Base Image | 10:44 |
| dmellado | 2021-12-20 10:43:49.951 | % Total % Received % Xferd Average Speed Time Time Time Current | 10:44 |
| dmellado | 2021-12-20 10:43:49.951 | Dload Upload Total Spent Left Speed | 10:44 |
| dmellado | xD | 10:44 |
| dtantsur | it's never going to be not DNS | 10:45 |
| dmellado | anyway, progressing, so let's see | 10:45 |
| dmellado | yah, works | 10:46 |
| dmellado | some network issue, now that it's cached it'll probably work | 10:46 |
| rpittau | currently in CI with dib built images I'm seeing this error OSError: [Errno 101] ENETUNREACH but haven't had the time to investigate further | 10:46 |
| rpittau | this happens on both debian and centos | 10:47 |
| dmellado | does it happens on the dib build phase? | 10:47 |
| rpittau | no, during deployment, the image builds fine | 10:47 |
| rpittau | the patch is https://review.opendev.org/c/openstack/bifrost/+/821223 | 10:47 |
| rpittau | that fixes the issue with pip from source | 10:47 |
| * dmellado fetching | 10:48 | |
| dmellado | thanks! | 10:48 |
| dtantsur | folks, we don't have the weekly meeting today, do we? | 10:48 |
| rpittau | dtantsur: we don't | 10:48 |
| *** sshnaidm|afk is now known as sshnaidm | 11:07 | |
| dmellado | dtantsur: also, is there any quick var or setting to avoid using /root/ for the image creation? that partition is quite full | 11:15 |
| dtantsur | dmellado: I suspect we run DIB with sudo.. | 11:16 |
| dmellado | probably | 11:16 |
| dmellado | I guess I'll have to check on the dib if there is any setting to use $whatever path | 11:17 |
| dtantsur | yep | 11:17 |
| dmellado | on the good side, I'm working on another PR to improve a few things on the dnsmasq.j2 file xD | 11:17 |
| dtantsur | nice! | 11:18 |
| * dmellado thinking about being cheap and just symlinking /httpboot | 11:21 | |
| dmellado | xD | 11:21 |
| opendevreview | Merged openstack/bifrost master: Add support for root filesystem UUID customisation https://review.opendev.org/c/openstack/bifrost/+/819785 | 11:47 |
| dmellado | https://opendev.org/openstack/diskimage-builder/src/branch/master/diskimage_builder/elements/redhat-common/bin/extract-image#L46 | 12:27 |
| dmellado | hmm so do you happen to know whether you can modify the TEMP file from bifrost? | 12:28 |
| dtantsur | dmellado: just set TMP_DIR? | 12:28 |
| dmellado | nvm, I'M BLIND | 12:29 |
| dmellado | time for coffee | 12:29 |
| dmellado | xD | 12:29 |
| dtantsur | coffee++ | 12:30 |
| dmellado | caveat: my kid did wake me up at 6am | 12:30 |
| dtantsur | ouch :( | 12:35 |
| dmellado | 2021-12-20 12:45:04.605 | > Error: Unable to find a match: scsi-target-utils | 12:45 |
| dmellado | does this ring a bell when building the image? | 12:45 |
| dmellado | rpittau: does this happen on CI? | 12:45 |
| dmellado | Seems like an old bug.... | 12:46 |
| dmellado | https://bugs.launchpad.net/tripleo/+bug/1864427 | 12:46 |
| dmellado | is this a regression on you folks? | 12:49 |
| rpittau | Mmm I don't see it in ci | 12:49 |
| rpittau | The only issue in seeing is a network connection where the ipa ramdisk can't reach ironic | 12:49 |
| dmellado | https://paste.openstack.org/show/811769/ | 12:51 |
| dmellado | seems that tgt gets converted to scsi-image-utils then boom | 12:51 |
| rpittau | dmellado: I'll check after lunch, I'm on my mobile at the moment :) | 12:53 |
| dmellado | no worries, thanks in advance! | 12:53 |
| *** amoralej is now known as amoralej|lunch | 13:06 | |
| dtantsur | why are we even trying to install it, I wonder | 13:06 |
| rpittau | dmellado: that's from running dib directly ? | 13:22 |
| rpittau | from what I can see the problem is in dib, it doesn't take into account that there's no scsi-target-utils for cnetos 8 | 13:24 |
| rpittau | so it's the same bug | 13:27 |
| rpittau | I believe "deploy-tgtadm" should not be included anymore | 13:29 |
| holtgrewe | arne_wiebalck: is there a way to specify the target image size when using diskimage-builder based on an existing qcow2? My image build crashes like this https://pastebin.com/1Wf7B5ty | 13:35 |
| holtgrewe | block device config yaml is here https://paste.openstack.org/show/811771/ | 13:36 |
| dmellado | rpittau: both running dib or bifrost, but seems to somehow have worked the 2nd time, let me ehcck | 13:48 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic master: Automatically configure enabled_***_interfaces https://review.opendev.org/c/openstack/ironic/+/820909 | 14:11 |
| *** amoralej|lunch is now known as amoralej | 14:12 | |
| jamesdenton | hi dtantsur - do you know in what scenario (if any) where the inspection_network specified in ironic.conf is leveraged for in-band inspection? | 14:30 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic master: [WIP] ImageCache: respect Cache-Control: no-store https://review.opendev.org/c/openstack/ironic/+/822329 | 14:47 |
| dtantsur | jamesdenton: hi, in this: https://docs.openstack.org/ironic/latest/admin/inspection.html#managed-inspection | 14:49 |
| jamesdenton | thanks - is managed triggered with 'openstack baremetal node inspect' vs 'openstack baremetal introspection start'? Just trying to determine the difference there. I have both methods working, but trying the former is using the ilo inspection interface (which doesn't appear to use network) and the latter doesn't seem to manipulate networking | 14:52 |
| dmellado | rpittau: I don't really see how can modify that in dib from bifrost | 14:54 |
| dtantsur | jamesdenton: well, it's a bit more subtle | 14:54 |
| dtantsur | 'baremetal introspection start' is a low-level ironic-inspector command that always uses unmanaged inspection | 14:54 |
| jamesdenton | ok | 14:55 |
| dtantsur | 'baremetal node inspect' uses the configured inspection interface, which can be inspector | 14:55 |
| dtantsur | your node clearly has 'ilo' inspection interface configured, which is out-of-band inspection | 14:55 |
| dtantsur | what you need is $ baremetal node set --inspect-interface inspector <node> | 14:55 |
| dtantsur | or put it as default_inspect_interface to your ironic.conf | 14:55 |
| dtantsur | then `baremetal node inspect` will use ironic-inspector | 14:56 |
| dtantsur | jamesdenton: ^^ | 14:56 |
| rpittau | dmellado: you can't, that is embedded in dib https://opendev.org/openstack/diskimage-builder/src/branch/master/diskimage_builder/elements/ramdisk/element-deps | 14:57 |
| jamesdenton | gotcha. i did try that, but get a 'notsupported' error, like this: openstack.exceptions.NotSupported: The baremetal-introspection service for 10.20.0.11:None exists but does not have any supported versions. So i must be missing some configuration, or maybe a mismatch in interfaces | 14:57 |
| jamesdenton | thank you for the tips - i'll dig around on my side and see whats up | 14:59 |
| rpittau | dmellado: for centos8 that should probably be change into targetcli and iscsi-initiator-utils | 15:01 |
| dtantsur | jamesdenton: this means that ironic cannot talk to ironic-inspector. something misconfigured in keystone or in the [inspector] section. | 15:15 |
| jamesdenton | yes - i found ironic-inspector API wasn't UP due to a local issue with dnsmasq. I've fix that, thank you. | 15:16 |
| dmellado | rpittau: but where is that change supposed to be around? | 15:16 |
| dmellado | I assume in diskimagte-builder directly | 15:16 |
| dmellado | but not sure about where the change is | 15:16 |
| rpittau | dmellado: yeah, probably the best place would be the centos element | 15:17 |
| rpittau | dmellado: this is where the pkg tgt gets remapped for redhat-based systems https://opendev.org/openstack/diskimage-builder/src/branch/master/diskimage_builder/elements/redhat-common/bin/map-packages#L85 | 15:19 |
| jamesdenton | dtantsur do you know if networking-generic-switch is supported with inspection? | 15:20 |
| dmellado | Oh, I was there but the WARNING really scared me xD | 15:20 |
| dmellado | # This global list has been deprecated by the pkg-map element. New | 15:20 |
| dmellado | # package mappings should go in pkg-map files inside each element. | 15:20 |
| dtantsur | jamesdenton: should work IMO | 15:20 |
| rpittau | dmellado: yes, the correct remap is done using the pkg-map element | 15:20 |
| dtantsur | you'll need to make sure you have all the information configured on the ports in advance | 15:20 |
| jamesdenton | right, ok | 15:21 |
| rpittau | dmellado: probably here https://opendev.org/openstack/diskimage-builder/src/branch/master/diskimage_builder/elements/redhat-common/pkg-map or in the centos/centos-minimal element | 15:28 |
| dmellado | rpittau: I see | 15:32 |
| dmellado | I'll make a PR | 15:32 |
| dmellado | but then there's this thing which creeps me out | 15:32 |
| dmellado | also busybox is not there? | 15:32 |
| dmellado | 2021-12-20 15:27:42.762 | Running install-packages install. | 15:32 |
| dmellado | 2021-12-20 15:27:42.762 | + dnf -v -y install busybox | 15:32 |
| dmellado | 2021-12-20 15:27:42.888 | Loaded plugins: builddep, changelog, config-manager, copr, debug, debuginfo-install, download, generate_completion_cache, groups-manager, needs-restarting, playground, repoclosure, repodiff, repograph, repomanage, reposync | 15:32 |
| dmellado | 2021-12-20 15:27:42.888 | DNF version: 4.7.0 | 15:33 |
| dmellado | 2021-12-20 15:27:42.889 | cachedir: /tmp/yum | 15:33 |
| dmellado | 2021-12-20 15:27:42.896 | User-Agent: constructed: 'libdnf (CentOS Linux 8; generic; Linux.x86_64)' | 15:33 |
| dmellado | 2021-12-20 15:27:42.997 | repo: using cache for: appstream | 15:33 |
| dmellado | 2021-12-20 15:27:43.009 | appstream: using metadata from Fri Dec 10 13:03:48 2021. | 15:33 |
| dmellado | 2021-12-20 15:27:43.010 | repo: using cache for: baseos | 15:33 |
| dmellado | 2021-12-20 15:27:43.022 | baseos: using metadata from Fri Dec 10 13:03:32 2021. | 15:33 |
| dmellado | 2021-12-20 15:27:43.022 | repo: using cache for: extras | 15:33 |
| dmellado | 2021-12-20 15:27:43.022 | extras: using metadata from Fri Dec 10 12:56:25 2021. | 15:33 |
| dmellado | 2021-12-20 15:27:43.022 | Last metadata expiration check: 4:41:49 ago on Mon Dec 20 10:45:54 2021. | 15:33 |
| dmellado | 2021-12-20 15:27:43.194 | No match for argument: busybox | 15:33 |
| dmellado | 2021-12-20 15:27:43.199 | Error: Unable to find a match: busybox | 15:33 |
| dmellado | xD | 15:33 |
| dmellado | ooops, I meant to paste the paste | 15:33 |
| dmellado | not this | 15:33 |
| dmellado | https://paste.openstack.org/show/811773/ | 15:33 |
| dmellado | sorry, wrong paste xD | 15:33 |
| holtgrewe | arne_wiebalck: I'm now quite happy with my CentOS 7 image. The next step forward for me would be to make this work for rocky linux... This appears to be a bit more involved, and there already is a patch here https://review.opendev.org/c/openstack/diskimage-builder/+/805800 | 15:38 |
| holtgrewe | Sorry, is this even the right place to ask? | 15:38 |
| holtgrewe | It looks like the commenters have reservations against supporting RockyLinux for hosting openstack itself. | 15:39 |
| holtgrewe | All I want is to build a RockyLinux image for my baremetal compute hosts... | 15:39 |
| opendevreview | Merged openstack/ironic bugfix/19.0: Fix redfish update_firmware for newer Sushy https://review.opendev.org/c/openstack/ironic/+/822241 | 15:40 |
| dtantsur | holtgrewe: #openstack-dib is a somewhat better channel for this | 15:41 |
| dtantsur | I know TheJulia has +2 rights on DIB, but that's it | 15:41 |
| dmellado | no, no busybox on centos8 | 15:41 |
| dmellado | aweeesome | 15:41 |
| dmellado | xD | 15:41 |
| holtgrewe | dtantsur: thanks | 15:41 |
| rpittau | dmellado: you found out yourself :) | 15:41 |
| arne_wiebalck | holtgrewe: nice, congrats! | 15:41 |
| arne_wiebalck | holtgrewe: TFW when the instantiated node starts pinging is still great , even after all these instance creations :) | 15:42 |
| dmellado | is busybox *really* needed? xD | 15:42 |
| dtantsur | dmellado: I'm really curious how you ended up with busybox.. | 15:43 |
| dtantsur | what's the command you're running? | 15:43 |
| arne_wiebalck | holtgrewe: I think the comments are more on the how than on the why FWIS. None the less, adding it to DIB means maintenance work, so I guess the question is whether there is sufficient demand. | 15:44 |
| arne_wiebalck | holtgrewe: feel free to comment on the ticket that you would like/use it, I guess there will be others to use it eventually | 15:45 |
| dmellado | dtantsur: dib uses it for creating the centos8 image xD | 15:45 |
| dmellado | ramdisk-image-create -o /httpboot/deployment_image.qcow2 -t qcow2 centos vm enable-serial-console simple-init | 15:45 |
| dmellado | just that xD | 15:45 |
| holtgrewe | arne_wiebalck: I see. I was looking into using the centos element for installing rocky but it's missing python3... | 15:45 |
| dtantsur | dmellado: okay, don't use ramdisk-image-create | 15:46 |
| dtantsur | honestly, we may even ask them to remove it | 15:46 |
| dtantsur | *always* use disk-image-create | 15:46 |
| dmellado | ack, thought it was going to be faster | 15:46 |
| dmellado | switching to disk-image-builder now | 15:46 |
| dtantsur | holtgrewe: a completely different approach would be to ask Rocky folks to publish images with a UEFI partition and linux-firmware included :) | 15:47 |
| arne_wiebalck | dtantsur: ++ | 15:47 |
| dtantsur | DIB is cool, but it has a tiny team that maintains it | 15:48 |
| holtgrewe | dtantsur: push it up up upstream :-D | 15:48 |
| holtgrewe | In the worst case, I'll go for maintaing my own elements. | 15:48 |
| dtantsur | well, yes :) | 15:48 |
| dtantsur | yeah, that's an option as well | 15:48 |
| dtantsur | right now https://rockylinux.org/cloud-images is only about AWS. this is pretty bad :) | 15:49 |
| rpittau | I was looking at rocky linux as well but I stopped when I ended up there :) | 15:49 |
| dtantsur | aha, https://download.rockylinux.org/pub/rocky/8.5/images/ has a genericcloud image | 15:50 |
| dtantsur | I wish they linked to it somewhere | 15:50 |
| rpittau | and that became interesting again | 15:50 |
| holtgrewe | So far you (the OpenStack) team appear to be the most friendly and approachable folks... anyway, your tools look to be well extensible | 15:50 |
| holtgrewe | So I can just help myself. | 15:50 |
| dtantsur | I'm glad you have good experiences here :) | 15:51 |
| jamesdenton | thanks for the help dtantsur -- it took the right combination of things to get it going. Namely, using ipmi driver (not iLo) and (maybe) require_managed_boot=True | 15:52 |
| dtantsur | jamesdenton: sounds like a good start | 15:53 |
| dtantsur | you can look into the iLO driver if you need any advanced features from it | 15:53 |
| jamesdenton | i don't, really. I just was trying to avoid flat networking for inspector | 15:53 |
| jamesdenton | which, seems like it's possible with the right configuration :) | 15:54 |
| dtantsur | jamesdenton: I definitely tried it myself a couple of years ago (with networking-ansible, which is pretty similar to networking-generic-switch) | 15:54 |
| jamesdenton | trying to get OSA's documentation and workflow in place | 15:54 |
| jamesdenton | i just need to figure out why this node is picking up a rogue IP via DHCP - i think i've got some cruft to cleanup | 15:56 |
| jamesdenton | thanks again | 15:56 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic bugfix/19.0: Configure CI for bugfix/19.0 https://review.opendev.org/c/openstack/ironic/+/822337 | 15:58 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic-python-agent bugfix/8.3: Configure CI for bugfix/8.3 https://review.opendev.org/c/openstack/ironic-python-agent/+/822338 | 16:01 |
| rpittau | I'm starting to wonder if the issue with dibipa integration job in bifrost is the python version | 16:01 |
| dtantsur | mmm? | 16:02 |
| rpittau | we use debian in dibipa integration, with python 3.9 | 16:02 |
| rpittau | I just saw another error that may be related to that: Could not get real physical RAM from lshw: Expecting ',' delimiter: line 20 column 8 (char 477): json.decoder.JSONDecodeError: Expecting ',' delimiter: line 20 column 8 (char 477) | 16:03 |
| rpittau | in tinyipa that does not happen (python 3.6) | 16:03 |
| dtantsur | rpittau: sounds rather like something with lshw | 16:04 |
| rpittau | also the issue with the network unreachable is quite strange, I can reproduce it locally | 16:04 |
| rpittau | dtantsur: yeah could be some lshw change | 16:04 |
| dtantsur | # lshw -quiet -json | jq . | 16:05 |
| dtantsur | parse error: Expected separator between values at line 11, column 8 | 16:05 |
| dtantsur | it produces invalid JSON on bullseye..... | 16:05 |
| rpittau | yay ? | 16:05 |
| dtantsur | absolutely cursed | 16:05 |
| dtantsur | this is version 02.18.85-0.7 | 16:06 |
| rpittau | on fc35 we have 02.19 and it works fine | 16:07 |
| dtantsur | FYI zigo ^^^ | 16:07 |
| dtantsur | rpittau: I don't know what to suggest, to be honest. debian needs to fix their package. | 16:09 |
| rpittau | yeah, nothing we can really do there | 16:12 |
| rpittau | I'll go back bumping my head on the network issue | 16:12 |
| dtantsur | ++ | 16:12 |
| dtantsur | yeah, the output is complete garbage | 16:14 |
| holtgrewe | It looks to me that my machine is not properly cleaned but boots something else via DHCP... do I *need* a cleaning network or would that also work with flat network mode? | 16:14 |
| dtantsur | holtgrewe: you can run cleaning on the same network as everything else | 16:15 |
| dtantsur | Ironic needs to know your cleaning_network because it needs to create ports in Neutron while cleaning | 16:15 |
| holtgrewe | dtantsur: OK... then it must be something else. | 16:16 |
| dtantsur | holtgrewe: if you have a conflicting DHCP server, it may be an issue | 16:16 |
| dtantsur | I mean, DHCP is okay, but a PXE server is a no-go | 16:16 |
| holtgrewe | Well, IT is failure after failure until something succeeds and then back to failure ;-) | 16:16 |
| dtantsur | :D | 16:16 |
| arne_wiebalck | holtgrewe: there is also https://docs.openstack.org/ironic/latest/user/creating-images.html I just realised | 16:30 |
| *** amoralej is now known as amoralej|off | 16:33 | |
| dtantsur | rpittau: tried reporting it to the debian's tracker, I hope it didn't go to /dev/null :) | 16:35 |
| rpittau | dtantsur: fingers crossed :) | 16:36 |
| dtantsur | rpittau, zigo, https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1002025 | 16:37 |
| zigo | dtantsur: It's well known, however it's fixed in Unstable (and backports). | 16:52 |
| rpittau | see ya tomorrow folks! o/ | 17:03 |
| arne_wiebalck | rpittau: o/ | 17:05 |
| dtantsur | zigo: yeah, but we don't use unstable.. | 17:13 |
| dtantsur | anyway, see you tomorrow everyone | 17:13 |
| opendevreview | Merged openstack/ironic stable/xena: Fix redfish update_firmware for newer Sushy https://review.opendev.org/c/openstack/ironic/+/822242 | 17:19 |
| opendevreview | Merged openstack/ironic-tempest-plugin master: [trivial] add python 3.9 in classifier https://review.opendev.org/c/openstack/ironic-tempest-plugin/+/821664 | 17:19 |
| arne_wiebalck | bye everyone o/ | 17:26 |
| admiyo | Running two Kolla setups, one old and working and one new and not working. What is supposed to respond to a DHCP request when provisioning a node? I thought it was ironic-dnsmasq, but that seems to get an entry in /etc/dnsmasq/dhcp-hostsdir/0c:42:a1:49:df:4c that just reads 0c:42:a1:4a:2c:ec,ignore | 18:20 |
| admiyo | And that is the same on both clusters. So, on the working cluster, what process should then handle the DHCP request? | 18:21 |
| admiyo | nova_compute_ironic ? | 18:25 |
| stevebaker[m] | good morning | 19:56 |
| zigo | dtantsur: I tried to get the stable team to allow uploading the version of unstable in stable, but the problem is the new gtk3 build-dep ... :/ | 22:25 |
| zigo | dtantsur: Would it help if I upload to backports? | 22:25 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!