| mup | Bug #1770072 opened: exiting "failed to enter rescue mode" should (optionally) not reboot <MAAS:New> <https://launchpad.net/bugs/1770072> | 05:05 |
|---|---|---|
| tosaraja | i'm trying to deploy with maas-cli a custom debian i've uploaded. What would the value be in "distro_series"? No matter what i try there, it doesn't seem to find the custom uploaded image. Can't they be deployed via maas-cli? | 05:56 |
| mup | Bug #1770091 opened: Special filesystems added twice on deploy <MAAS:New for ack> <https://launchpad.net/bugs/1770091> | 06:59 |
| === frankban|afk is now known as frankban | ||
| tosaraja | and seems the raw file i downloaded from debian needs to be comprssed before uploading as maas expect it to be a tarball | 08:18 |
| tosaraja | (and still have to deploy it from the gui) | 08:18 |
| tosaraja | but next problem in my hands is that debian doesn't have dpkg and maas fails deployment | 08:18 |
| roaksoax | tosaraja: debian is not suported | 08:21 |
| tosaraja | roaksoax: that's it then... guessi have to go back to trying centos | 08:21 |
| roaksoax | tosaraja: and if your image doesn't have dpkg, that's a problem in the image, not MAAS | 08:22 |
| tosaraja | yeah, agreed there | 08:22 |
| tosaraja | tried just downloading the stock raw image from debian if would have worked. | 08:22 |
| roaksoax | that's strange though, that the image doesn't have dpkg in it | 08:23 |
| tosaraja | aaah not necessarily btw | 08:23 |
| tosaraja | Stderr: chroot: failed to run command 'dpkg': No such file or directory | 08:23 |
| tosaraja | could that error come from dpkg not finding what it is looking for | 08:23 |
| tosaraja | yeah the command is most likely unshare', '--fork', '--pid', '--', 'chroot', '/tmp/tmp_li3uyg8/target', 'dpkg', '--print-architecture | 08:24 |
| tosaraja | so it is dpkg it can't find | 08:25 |
| roaksoax | yeah, no idea unfortunately. I've never tried to deploy debian | 08:28 |
| roaksoax | why not just use ubuntu ? | 08:28 |
| tosaraja | they tend to crash too often. either mysterious errors while forking a process or nfscache just hitting a known bug that hasn't been fixed. We've tried using 16.04, 17.04, 17.10 and now we should try 18.04, but it didn't start deploying right of the bat so to say | 08:32 |
| tosaraja | saw that someone had figured out how to upload debian into maas... yeah that went ok after a few hours of figuring out the tarball issue and the fact that it the maas-cli apparently doesn't know how to deploy them. i guess it's time to pick the fight between centos 7 and 18.04 now | 08:33 |
| roaksoax | tosaraja: right, but if you having those nfs related crashes in Ubuntu you'll most likely have them in debian as well | 08:33 |
| tosaraja | indeed possible | 08:34 |
| bjarne | Hi, does anyone know how i can list all hosts in a domain using cli, the gui eqivalent of :5240/MAAS/#/domain/0 | 08:38 |
| roaksoax | bjarne: maas <user> domains read ? | 08:42 |
| roaksoax | bjarne: maas <user> domain read <domain> # for a single domain | 08:43 |
| bjarne | Welll, I want all hosts in that domain, not the domain info it self. The gui says "maas (default): 59 hosts, 135 records total" with records like "admin.compute-1 A 192.168.x.x" Its the records i want | 09:00 |
| mup | Bug #1770126 opened: [2.4b3] Upgrading a rack controller doesn't automatically update image status or version <MAAS:Triaged by blake-rouse> <https://launchpad.net/bugs/1770126> | 09:05 |
| mup | Bug #1770130 opened: [2.4] API doesn't list dns resources automatically created by MAAS <MAAS:Triaged by mpontillo> <https://launchpad.net/bugs/1770130> | 09:21 |
| roaksoax | bjarne: /win 4 | 09:42 |
| tosaraja | If I choose to deploy a CentOS7, why is it loading "error: no such device: /efi/ubuntu/shimx64.efi." that says ubuntu, and the grub is also Ubuntus? | 09:55 |
| roaksoax | tosaraja: https://bugs.launchpad.net/maas/+bug/1740925 | 10:17 |
| tosaraja | ah found the problem: https://bugs.launchpad.net/maas/+bug/1740925 | 10:17 |
| roaksoax | indeed | 10:18 |
| tosaraja | lol, what timing :D | 10:18 |
| tosaraja | after 10 minutes of searching we paste the link within 2 seconds :D' | 10:18 |
| roaksoax | tosaraja: http://paste.ubuntu.com/p/kRyT5pvyFR/ /usr/lib/python3/dist-packages/provisioningserver/templates/uefi/config.local.amd64.template | 10:21 |
| roaksoax | tosaraja: that /may/ fix the issue | 10:21 |
| tosaraja | thanks, i'll have a look | 10:21 |
| roaksoax | tosaraja: you'd need to restart maas-rackd though | 10:21 |
| roaksoax | tosaraja: if that works, I can commit that | 10:21 |
| tosaraja | or could i just update to maas 2.4.0beta4? | 10:21 |
| tosaraja | ah, that's the fix apparently | 10:21 |
| roaksoax | i have not tested that, but that may fix it | 10:22 |
| tosaraja | i can test it | 10:22 |
| bjarne | roaksoax: what do you mean "/win 4" | 10:25 |
| tosaraja | roaksoax: modified the file under /usr/lib/python3/dist-packages/provisioningserver/templates/uefi | 10:33 |
| tosaraja | let's see if it affected anything | 10:34 |
| tosaraja | roaksoax: it did what we'd expect it to do...kinda :D error: no such device: /efi/redhat/grub.efi. | 10:39 |
| mup | Bug #1770141 opened: [2.4b3] When SMART is not supported on a drive, some tests pass and some are skipped <MAAS:New> <https://launchpad.net/bugs/1770141> | 11:15 |
| roaksoax | bjarne: sorry, that was a typo | 11:25 |
| bjarne | ok, so no idea how to get the hosts for a domain then? | 11:35 |
| bjarne | im on 2.3.0 | 11:36 |
| roaksoax | bjarne: there's an open bug where the api doesn't show the hostnames under dnsresources for those that maas automatically genreated | 11:40 |
| roaksoax | i thought that had been fixed | 11:40 |
| roaksoax | but obviously it hasnt | 11:40 |
| KingJ_ | What's the recommended way of getting MAAS's DNS entries in to an upstream DNS server? | 12:47 |
| roaksoax | KingJ_: maas doesn't send entries into an upstream DNS server | 12:48 |
| roaksoax | KingJ_: maas can be configured to use another dns server as a forwarder | 12:48 |
| === KingJ_ is now known as KingJ | ||
| roaksoax | KingJ_: but you could tweak your dns server to send stuff uip to another if you dont mess with maas' configuration I guess | 12:48 |
| KingJ | Indeed, which lets MAAS resolve for its own managed servers and other items upstream, but what i'm looking to do is allow non-MASS managed hosts to resolve MASS managed ones - so somehow I need to either get MAAS's DNS entries on to an upstream DNS server, or have the upstream DNS server query MAAS for the specific MAAS domain? | 12:49 |
| roaksoax | KingJ: i dont specifically know. I've never done anything similar myself | 12:51 |
| roaksoax | although, it may be possible by configuring bind to do so | 12:51 |
| KingJ | I'm a little adverse to modifying MASS's own bind configuration as it may get wiped out on an upgrade... but equally i'm also deploying MAAS hosts in to an existing domain rather than a MASS specific domain, so I can't get my upstream DNS server to forward e.g. requests for host.mass.domain.com to the MAAS server. | 12:52 |
| KingJ | The only other way i've possibly seen mentioned is adding a DHCP snippet that uses Dynamic DNS to register the hostname in my upstream DNS server? But that feels iffy... having the host register itself with my upstream DNS server rather than having MAAS push that through? | 12:53 |
| roaksoax | KingJ: yeah it feels iffy to me too | 12:54 |
| roaksoax | i would not do DDNS | 12:54 |
| roaksoax | although its possible | 12:54 |
| roaksoax | i think thta's probably the only way to do it | 12:54 |
| KingJ | I've just stumbled across this too, which hmm, could work - get the upstream DNS server to do a zone transfer. Although i'm not sure if that'd work given it's the same zone (plus, it's not yet implemented so for now I would need to alter the config, but I feel like that's less hacky than the DHCP snippet for DDNS?) | 12:56 |
| KingJ | https://bugs.launchpad.net/maas/+bug/1711395 | 12:56 |
| mup | Bug #1770201 opened: During enlistment DNS resolution failures leave required packages uninstalled <cdo-qa> <cdo-release-blocker> <foundations-engine> <MAAS:New> <https://launchpad.net/bugs/1770201> | 16:13 |
| === frankban is now known as frankban|afk | ||
| mup | Bug #1732194 changed: [2.3rc2, UI] When I want to add a new interface to a device the MAC address field has a red border <2.3qa> <ui> <ui-papercut> <MAAS:Invalid by m-vrachnis> <https://launchpad.net/bugs/1732194> | 16:43 |
| mup | Bug # changed: 1764830, 1769448, 1769468, 1769471 | 18:25 |
| mup | Bug # opened: 1764830, 1769448, 1769468, 1769471 | 18:28 |
| mup | Bug # changed: 1764830, 1769448, 1769468, 1769471 | 18:34 |
| maticue | Hi there, trying to find Supermicro 1028U-TNRTP+ using MAAS 2.3.0. MAAS discovers the baremetal but it is unable to configure IPMI automatically. In fact, it doesn't work setting it manually. I will paste the log error. Seems to be an IPMI communication issue | 19:52 |
| maticue | May 9 19:30:30 sea-ec220-mrk-1 maas.power: [info] Changing power state (on) of node: cuddly-iguana (nyamym) May 9 19:30:30 sea-ec220-mrk-1 maas.power: [error] Error changing power state (on) of node: cuddly-iguana (nyamym) May 9 19:34:12 sea-ec220-mrk-1 maas.power: [error] cuddly-iguana: Power state could not be queried: expect bytes or str, not NoneType May 9 19:34:13 sea-ec220-mrk-1 maas.power: [error] cuddly-iguana: Failed to r | 19:53 |
| roaksoax | maticue: what version of MAAS are you using ? | 19:56 |
| maticue | roaksoax: MAAS version: 2.3.0 (6434-gd354690-0ubuntu1~16.04.1) | 19:56 |
| roaksoax | maticue: have you tried upgrading to 2.3.2 and seeing if the issue is fixed ? | 20:03 |
| maticue | roaksoax: nop, but I can test it | 20:03 |
| maticue | roaksoax: I will do it and back to you | 20:03 |
| roaksoax | maticue: ok, that said, never seen that specific error before... | 20:21 |
| blahrus | Just setup MAAS on 18.04 - got servers PXE'ing off of it. | 21:08 |
| blahrus | We see the 2 machines under USED for the IP addresses | 21:09 |
| blahrus | but they don't show up under machines or Network discovery | 21:09 |
| blahrus | any ideas? | 21:09 |
| maticue | roaksoax: Solved! maas region controller was 2.3.0, maas rack controller was 2.2.2-6099-g8751f91-0ubuntu1~16.04.1. Upgraded maas rack controller to 2.3.0 and it worked. Thanks!!! | 21:10 |
| roaksoax | maticue: cool | 21:11 |
| roaksoax | blahrus: probably an issue we tracking where is not correctly running internal services | 21:12 |
| blahrus | roaksoax: Is there a fix? | 21:12 |
| roaksoax | blahrus rc1 will hvae lots of fixes improviding various backend processing issues | 21:13 |
| roaksoax | which would be related to weird behavior | 21:13 |
| roaksoax | like the one you describe | 21:13 |
| blahrus | ETA for that, or should I downgrade versions? | 21:15 |
| roaksoax | blahrus: hopefully friday | 21:16 |
| roaksoax | blahrus: but that siad, if it shows 2 ip's address | 21:18 |
| roaksoax | blahrus: it could also be that those are just lingering from the DHC{ and have not yet expired | 21:18 |
| blahrus | It shows the 2 systems got leases | 21:18 |
| blahrus | I watched cloud-init run on the boxes | 21:18 |
| blahrus | then they shut down | 21:18 |
| roaksoax | blahrus: well it could be the leases for other interface son the same nodes | 21:19 |
| roaksoax | e.g. during commissioning it tries to identify whether other interfaces are in the same l2 | 21:19 |
| blahrus | Right, but if got all the way through cloud-init and it shuts down - shouldn't somnething show up in Machines? | 21:20 |
| roaksoax | blahrus: it could have failed | 21:21 |
| roaksoax | blahrus: where you enlisting the machine ? | 21:21 |
| blahrus | Only thing I did was reboot the server to the PXE | 21:21 |
| mup | Bug #1706696 changed: maas auto-allocates machines via pods even when zone is included in constraints <cdo-qa> <internal> <MAAS:Invalid> <https://launchpad.net/bugs/1706696> | 21:26 |
| blahrus | roaksoax, looks like if I manually add the system and PXE boot it, MAAS sees it and sets it to ready | 21:35 |
| pines | I am trying to commission multiple machines with the CLI. I tried this: /usr/bin/maas root machines accept machines=[m1,m2,m3] but it is not working | 21:37 |
| pines | is there a way to send a list or something similar through the maas CLI for multiple machines? | 21:37 |
| mup | Bug #1770273 opened: [2.4, UI] IP assignment form for a machine has been mixed with that one of a device <MAAS:Triaged by deadlight> <https://launchpad.net/bugs/1770273> | 21:41 |
| roaksoax | pines: maas admin machines accept machines=8kngxk machines=htbkpm | 21:43 |
| roaksoax | pines: but oyu need to use the system-ids | 21:44 |
| roaksoax | pines: or you can do accept-all iirc | 21:44 |
| roaksoax | maas <user> machines accept machines=<sustem_id> machines=<syustem_id> | 21:44 |
| pines | roaksoax: Thanks a lot... I thought I could sort it out with a list, but now it is clear to me. It is confusing if it sasy machines..should say machine, so you know it only allows one :) | 21:47 |
| mup | Bug #1770273 changed: [2.4, UI] IP assignment form for a machine has been mixed with that one of a device <MAAS:Triaged by deadlight> <https://launchpad.net/bugs/1770273> | 21:50 |
| roaksoax | pines: yeah we are hoepfully redesisgning the cli next cycle | 21:52 |
| mup | Bug #1770273 opened: [2.4, UI] IP assignment form for a machine has been mixed with that one of a device <MAAS:Triaged by deadlight> <https://launchpad.net/bugs/1770273> | 21:56 |
| pines | roaksoax: Would it be the same case for teh following command? /usr/bin/maas root machines set-zone nodes=system_id1 nodes=system_id2 zone=ZONE | 22:13 |
| pines | or I need to do this one by one? | 22:13 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!