| mup | Bug #1777967 opened: dpkg-query: package 'maas-cluster-controller' is not installed and no inn is available <maas-maintainers> <MAAS:New> <https://launchpad.net/bugs/1777967> | 02:02 |
|---|---|---|
| === mup_ is now known as mup | ||
| === frankban|afk is now known as frankban | ||
| naturalblue | hi. I am trying to deploy a juju maas controller and it starts to deploy but i notice later on it looks for an IP (that was MAAS's old IP address) and then fails to deploy. Is there a way to get the details of what it is looking for. I have changed all ips and restarted all services on MAAS. I am not sure why it is still looking for the old address. | 11:05 |
| naturalblue | would puttting the node into rescue mode and then logging in allow me to see the error | 11:05 |
| mup | Bug #1778047 opened: No rack controllers can access the BMC of node <MAAS:New> <https://launchpad.net/bugs/1778047> | 11:48 |
| naturalblue | i thought if you uploaded your ssh public key to maas then it was compied over to all machines it deploys so you can login. is this only on rescue mode? | 14:04 |
| PatrickD_ | not only rescue mode. | 14:33 |
| PatrickD_ | on commissioned and deployed | 14:33 |
| mup | Bug #1777967 changed: dpkg-query: package 'maas-cluster-controller' is not installed and no inn is available <maas-maintainers> <MAAS:Invalid> <https://launchpad.net/bugs/1777967> | 15:03 |
| PatrickD_ | Having problems with Dell 640, with a Megaraid (perc) controller. fails to run smartctl (I guess it should either detect that it should use "-d megaraid,<diskid>" to scan the disks, or simply return that smartctl is not supported... | 15:07 |
| PatrickD_ | this is at commissioning, when it runs maas-run-remote-scripts | 15:08 |
| PatrickD_ | {"1.0": {"testing_scripts": [{"name": "smartctl-validate", "path": "testing/smartctl-validate", "script_result_id": 89, "script_version_id": 1, "timeout_seconds": 300, "parallel": 1, "hardware_type": 3, "parameters": {"storage": {"type": "storage", "value": "all"}}, "packages": {"apt": ["smartmontools"]}, "for_hardware": [], "has_started": true}]}} | 15:09 |
| mup | Bug #1777967 opened: dpkg-query: package 'maas-cluster-controller' is not installed and no inn is available <maas-maintainers> <MAAS:Invalid> <https://launchpad.net/bugs/1777967> | 15:12 |
| mup | Bug #1777967 changed: dpkg-query: package 'maas-cluster-controller' is not installed and no inn is available <maas-maintainers> <MAAS:Invalid> <https://launchpad.net/bugs/1777967> | 15:24 |
| === frankban is now known as frankban|afk | ||
| polto__ | hi | 17:38 |
| polto__ | is MAAS packaged with the latest LTS ubuntu (18.04) is considered stable ? | 17:38 |
| polto__ | I am facing errors trying to deploy servers from MAAS | 17:40 |
| roaksoax | polto__: 2.4b2 is in the archive | 17:40 |
| roaksoax | polto__: 2.4 final is on a ppa | 17:41 |
| PatrickD_ | @roaksoax Did you get a chance to read my question from 2 hours ago ? We have an issue with smartctl. We are digging, trying to find how the index.json is generated. ours seem to be missing stuff in the type:storage value: all... Should be a table there instead of "all". | 17:44 |
| ltrager | PatrickD_: MAAS uses 'all' as a place holder when a new machine is added. | 17:46 |
| ltrager | PatrickD_: What happens is the machine starts commissioning, sends commissioning results to MAAS, then maas-run-remote-scripts redownloads the tar which has the discovered values | 17:47 |
| ltrager | PatrickD_: when smartctl failed did commissioning successfully complete and discover disks? | 17:47 |
| PatrickD_ | no. It sees the size of the disk (/dev/sda, 1 single disk because it is a RAID5 on a Dell server PERC controller). and then fails with this error : https://paste.ubuntu.com/p/6CsfwRztds/ | 17:50 |
| ltrager | PatrickD_: Whats the output of the API for the machine? - maas $PROFILE machine read $SYSTEM_ID | 17:53 |
| PatrickD_ | https://paste.ubuntu.com/p/wwtQWHhcbK/ | 17:58 |
| PatrickD_ | maas-run-remote-scripts downloads the index.json, right ? And this json should not contain "all" in that spot. It should contain a hash. So maas-run-remote-scripts fails because .get method on a non-has fails ? | 18:00 |
| ltrager | PatrickD_: yes maas-run-remote-scripts downloads index.json in a tarball with all scripts that are pending | 18:02 |
| ltrager | PatrickD_: the "all" placeholder is fine during commissioning but not during testing | 18:02 |
| ltrager | PatrickD_: What should be happening is commissioning finishes, maas-run-remote-scripts redownloads the tar containing index.json and executes the tests | 18:02 |
| ltrager | PatrickD_: what version of MAAS are you running | 18:03 |
| PatrickD_ | 2.4 from ppa | 18:05 |
| PatrickD_ | so what script gathers the information necessary to build the proper storage information ? | 18:07 |
| ltrager | PatrickD_: The builtin commissioning script 00-maas-07-block-devices | 18:08 |
| PatrickD_ | https://paste.ubuntu.com/p/q9cj2sQd5g/ | 18:09 |
| PatrickD_ | result from the 00-maas-07-block-devices script | 18:09 |
| ltrager | PatrickD_: is there anything in the machines event log? | 18:15 |
| ltrager | PatrickD_: What should be happening is 00-maas-07-block-devices is received which causes MAAS to regenerate storage parameters. | 18:16 |
| ltrager | PatrickD_: That should replace 'all' with the actual data | 18:16 |
| PatrickD_ | event log ? | 18:20 |
| ltrager | PatrickD_: In the UI select the machine and click on the Events tab | 18:21 |
| PatrickD_ | Nothing special there... PXE Request - commissioning, then Node changed status - From 'Commissioning' to 'Testing' | 18:22 |
| PatrickD_ | then to Node changed status - From 'Testing' to 'Failed testing' | 18:22 |
| PatrickD_ | you think the output from the 07-block-devices script is fine ? (I haven't seen any outputs to really compare to :) | 18:23 |
| ltrager | PatrickD_: yeah it seems fine and the output from the API shows the disk is being added | 18:24 |
| ltrager | PatrickD_: could you file a bug at https://bugs.launchpad.net/maas/+filebug and include the logs from /var/log/maas? | 18:24 |
| ltrager | PatrickD_: to get around this bug you can commission without running any tests. In the UI you can just deselect smartctl-validate | 18:25 |
| ltrager | PatrickD_: Using the API you can change what runs by default by running | 18:26 |
| ltrager | PatrickD_: maas $PROFILE node-script remove-tag smartctl-validate tag=commissioning | 18:26 |
| PatrickD_ | Ok. Trying a few things, then I will file the bug with more info. | 18:41 |
| PatrickD_ | the command to remove smartctl is the same as doing it per machine in the UI ? | 18:44 |
| ltrager | PatrickD_: The API command stop its from being selected by default | 19:58 |
| ltrager | PatrickD_: This way if you use something like JuJu to commission it won't run | 19:58 |
| jaypipes | hi folks. I'm hoping a MaaS expert might be able to take a look at https://bugs.launchpad.net/maas/+bug/1778047 and provide some advice as to what I should try to get my machines able to be commissioned by the rack controller. thx in advance! | 21:09 |
| arkaroo | I can't seem to get maas to comission nodes. I have my setup at the point where nodes will PXEboot from the maas region controller, but no matter what I do, the PXEboot ends with the error "mounting /root on /media/root-ro failed: Invalid argument" followed by a bunch of other errors about not being able to mount to /root, and then just the busybox initramfs terminal prompt | 21:36 |
| arkaroo | it looks like squashfs is trying to mount to a path that doesn't exist, but I can't figure out why, anyone know what's going on? | 21:37 |
| arkaroo | for the record: it looks like if the PXEboot can't download the image it wants, it doesn't stop and tell you the network is incorrectly configured, it tries to extract a zip that isn't there and gives you mounting errors | 22:05 |
| arkaroo | not sure yet why I can pxeboot from the region controller but can't download files from it, something is very weird there | 22:06 |
| Wilee | Hello | 23:38 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!