| === frankban|afk is now known as frankban | ||
| mup | Bug #1717189 opened: Commision runs lsblk wit hunknown option <MAAS:New> <https://launchpad.net/bugs/1717189> | 07:37 |
|---|---|---|
| mup | Bug #1717176 opened: Fail to install or commission wichita (power 8) <MAAS:New> <curtin (Ubuntu):Invalid> <https://launchpad.net/bugs/1717176> | 07:55 |
| mup | Bug #1717176 changed: Fail to install or commission wichita (power 8) <MAAS:New> <curtin (Ubuntu):Invalid> <https://launchpad.net/bugs/1717176> | 08:04 |
| mup | Bug #1717176 opened: Fail to install or commission wichita (power 8) <MAAS:New> <curtin (Ubuntu):Invalid> <https://launchpad.net/bugs/1717176> | 08:07 |
| bryanb229 | hello wondering if anyone has successfully been able to deploy centos 7 with an out of the box maas installation | 09:58 |
| bryanb229 | ive been trying for many months but it keeps hanging on the web ui | 09:59 |
| bryanb229 | as it relates to this bug https://bugs.launchpad.net/maas/+bug/1627441 | 09:59 |
| mwe1 | @bryanb229: if your servername is 'localhost' after install: can you ensure there is no network related issue? DNS, etc... | 12:25 |
| bryanb229 | @mwe1 i dont think its a dns issue b/c centos 6 and ubuntu installs work fine, its just centos 7 that is not working | 12:30 |
| jamesbenson | rharper: ping | 14:09 |
| mup | Bug #1717269 opened: [Documentation] Intro concepts page issues <MAAS:New> <https://launchpad.net/bugs/1717269> | 14:38 |
| jamesbenson | rharper: ping | 15:47 |
| rharper | jamesbenson: here | 15:47 |
| jamesbenson | hey, have time for a quick chat about bug 1717031? | 15:47 |
| jamesbenson | I booted back into my maas v2.1.5 node to recommission this machine, but still no luck with uefi. | 15:49 |
| jamesbenson | I'm going to go back to v2.2.2. In regards to your comment about BIOS booting with GPT. We need to create that partition? Because I've never seen it. | 15:50 |
| rharper | sure | 15:50 |
| rharper | You shouldn't need to make changes manually, maas should know how to do GPT partitions for large drives; | 15:50 |
| jamesbenson | okay | 15:50 |
| rharper | I'm not familiar with the maas UI on what might need to change | 15:51 |
| jamesbenson | with bios, it detects the gpt, but there is only the one partition / | 15:51 |
| rharper | but generally, if MAAS knows it's GPT, it should automatically create the config to include the 1MB linux-bios partition (and set the flag) | 15:51 |
| rharper | it might not show you it in the UI; but again I'm not for sure (I work on the curtin installer which consumes the config that maas sends) | 15:52 |
| jamesbenson | regardless of eufi or bios? | 15:52 |
| jamesbenson | okay | 15:52 |
| jamesbenson | do you know the cli off hand to verify? | 15:52 |
| rharper | for UEFI, it has a /boot/EFI partition instead of a linux bios partition | 15:52 |
| rharper | yeah, let's run that command to extract the curtin config | 15:52 |
| catbus | need to recommission the machine if bios boot mode is changed, so maas will set the partition accordingly. | 15:53 |
| rharper | maas <session> machine get-curtin-config <system-id> | 15:53 |
| rharper | I think that should dump the config; and I can help verify | 15:53 |
| jamesbenson | okay, thanks guys. I just switched it back to bios. | 15:55 |
| jamesbenson | I'll recommission now and get back to you in a few when it is finished. | 15:55 |
| rharper | jamesbenson: ok | 15:57 |
| jamesbenson | looks like that partition is there: https://pastebin.com/qZJphMhY I'm deploying now to see how it goes. | 16:16 |
| alopez | hi, i need add a node that has already SO installed and has partition into maas server | 16:23 |
| alopez | is this possible ? | 16:24 |
| jamesbenson | rharper & catbus: installation was "successful" however, boot goes into grub rescue. | 16:29 |
| rharper | jamesbenson: what release are you installing ? Xenial ? | 16:29 |
| jamesbenson | maas v2.2.2 and 16.04 LTS | 16:30 |
| catbus | what's the error before it drops to grub rescue? | 16:30 |
| jamesbenson | Kernel: xenial (ga-16.04) OS: Ubuntu 16.04 LTS "Xenial Xerus" | 16:31 |
| jamesbenson | error: attempt to read or write outside of disk 'hd0' | 16:31 |
| rharper | interesting | 16:31 |
| alopez | anyone knows how to do this ? | 16:32 |
| catbus | jamesbenson: the boot disk drive is >2TB I assume? | 16:34 |
| === frankban is now known as frankban|afk | ||
| jamesbenson | yes: Storage 7999.4GB over 1 disks | 16:34 |
| jamesbenson | It does pick up that it is a PERC 6/i card... | 16:35 |
| rharper | what sort of raid mode? I don't think it matters; and it should work | 16:35 |
| jamesbenson | Raid 6 | 16:36 |
| rharper | but the error message seems to imply the sector where the grub config is beyond the grub limit; just looking around for more information | 16:36 |
| catbus | jamesbenson: could you try this? Create a partition for '/' with 'ext4' under sda in the node page, re deploy. | 16:36 |
| jamesbenson | Found initrd image: /boot/initrd.img-4.4.0-93-generic | 16:36 |
| jamesbenson | /run/lvm/lvmetad.socket: connect failed: No such file or directory | 16:36 |
| jamesbenson | WARNING: Failed to connect to lvmetad. Falling back to internal scanning. | 16:36 |
| jamesbenson | done | 16:36 |
| jamesbenson | In the installation logs it shows: | 16:36 |
| catbus | IIRC, the bios boot mode supports up to 2TB. | 16:36 |
| jamesbenson | catbus: that's what it is currently | 16:36 |
| catbus | jamesbenson: sorry, set the partition size to 1TB. | 16:36 |
| jamesbenson | ok | 16:37 |
| jamesbenson | releasing and re-deploying with that partition. | 16:37 |
| catbus | uefi boot mode is designed to support drive >2TB, the limitation of the legacy bios mode. | 16:38 |
| rharper | GPT handles >2TB fine ; I don't think the size is the issue here | 16:40 |
| jamesbenson | catbus: we have always used bios and been able to use the full drive without issues.... It may not have been only one partition, not sure. But maas doesn't even deploy with uefi. But I would love to get this figured out.... | 16:41 |
| rharper | we could create a /boot partition ; that might ensure that the /boot directory location on the root partition isn't too far | 16:41 |
| jamesbenson | rharper, that's what I did initially | 16:41 |
| jamesbenson | in the bug report. | 16:41 |
| jamesbenson | That's what I did in v2.1.5 to make it work | 16:41 |
| jamesbenson | 1GB /boot; 4GB SWAP; 8TB / | 16:41 |
| jamesbenson | but in 2.2.2 this doesn't wokr | 16:41 |
| jamesbenson | work | 16:41 |
| rharper | y; the original failure looked to be related to maas not knowing which boot mode you were in (UEFI vs. BIOS) | 16:42 |
| jamesbenson | No, I was attempting all methods that "should work", none did. | 16:43 |
| jamesbenson | :-( | 16:43 |
| rharper | that seems to been in sync now after commissionion; the UEFI shimx64 missing looked to be related to maybe secure-boot being enabled, but I don't know if it's only an issue with SB; the grub2 package install should do the right thing and ensure shim is installed | 16:43 |
| jamesbenson | We don't use secure boot. But I will double check that it is disabled. Currently I'm trying to deploy with the 1TB / with BIOS. | 16:45 |
| jamesbenson | FWIW: I'm willing to have someone remote login/view as well to check things out. | 16:47 |
| catbus | Let's see if 1TB / with BIOS works first. | 16:48 |
| catbus | UEFI should just work without any change on the partition table configured by MAAS after recommissioning. Or you may check if there is newer firmeware/BIOS. | 16:49 |
| jamesbenson | firmware/bios is latest on every piece of hardware.... I spent way too much time doing that... | 16:51 |
| catbus | ok | 16:51 |
| catbus | jamesbenson: is your server one of the 16.04 certified models here: https://certification.ubuntu.com/server/? | 16:52 |
| jamesbenson | .... no. | 16:53 |
| jamesbenson | we have dell r410, r610, r710, r910... this one is a r710. | 16:54 |
| jamesbenson | I see the r410 & r610 are there, but not the 710&910.... | 16:54 |
| jamesbenson | I see it is compatible with 14.04 | 16:55 |
| catbus | ok. | 16:56 |
| jamesbenson | and our raid controller is not on that list for the r710 | 16:56 |
| catbus | Is the installation done yet? | 16:57 |
| jamesbenson | sorry I got stuck with the secure boot to see if it is enabled. It's not available when you're in BIOS, needs to be UEFI. Going back to deployment. | 17:03 |
| jamesbenson | maas did say the curtin was finished. | 17:04 |
| jamesbenson | I think it worked, but I'm having issues SSH'ing into the machine... | 17:17 |
| jamesbenson | success with the 1TB / | 17:18 |
| jamesbenson | catbus: ^^ | 17:18 |
| rharper | jamesbenson: \o/ | 17:30 |
| jamesbenson | so the question then, where is the bug? | 17:41 |
| jamesbenson | catbus, rharper: idea's? | 17:42 |
| catbus | jamesbenson: did the server boot up fine? | 17:43 |
| jamesbenson | yes with the 1TB / | 17:44 |
| catbus | jamesbenson: I assume that you have tried UEFI boot mode, recommission, and deploy, and it failed, what's the failure? | 17:46 |
| catbus | jamesbenson: BIOS (Legacy) boot mode uses MBR which doesn't support >2TB boot drive. UEFI boot mode uses GPT which supports boot drive >2TB. | 17:47 |
| jamesbenson | yes, I've tried UEFI boot following that procedure. The failure is: error: no such device: /efi/ubuntu/shimx64.efi | 17:48 |
| jamesbenson | catbus: installation is successful, on reboot that error occurs | 17:48 |
| jamesbenson | error: File not found. | 17:48 |
| rharper | the UEFI-mode missing shim is a grub2 bug; it's not yet know what the root cause for grub2 installation failing to install the shimx64 file; that needs following up; in your case, you're not using SecureBoot, that's an additional data point for that bug | 17:54 |
| jamesbenson | I'm enable UEFI and double check, but we shouldn't be using it. | 17:55 |
| rharper | w.r.t BIOS and 8TB drive, I think the /boot partition (rather than just a smaller than 2TB /) should work; the message from grub about read failures is likely due to the location of where grub configuration files ended up on the disk; a smaller dedicated /boot in addition to the / partition should work; | 17:56 |
| jamesbenson | s/I'm/I'll/ | 17:56 |
| jamesbenson | I'll try that again. I tried it before and it failed, but who knows. I'll try again. | 17:56 |
| catbus | agreed re: a smaller dedicated /boot in addition to the / partition should work. | 17:58 |
| jamesbenson | 1GB /boot should be more than enough correct? | 17:58 |
| rharper | jamesbenson: ok; I think we'll collect the data from today an update the bug, we can check in with the grub/efi folks to see if something else is going one (and if they have any steps to try once in the failing state to help isolate where the failure is; say firmware or grub2 ,etc) | 17:59 |
| rharper | yeah | 17:59 |
| jamesbenson | sounds good. I'm keeping v2.2.2 so feel free to message me if you need data. Fixing this bug helps the community and me. :-) | 18:00 |
| jamesbenson | rharper catbus: just deployed with BIOS, 1GB /boot 8GB SWAP 8TB / and it somehow worked. doing a fdisk -l, I can see a 1M BIOS boot partition there as well. | 18:17 |
| rharper | \o/ | 18:19 |
| mup | Bug #1717348 opened: Query iLO/iDRAC/BMC hostname/servername to set node name during enlisting instead of generating $random-$random name <MAAS:New> <https://launchpad.net/bugs/1717348> | 20:05 |
| mup | Bug #1717287 opened: maas-enlist doesn't work when provided with serverurl with IPv6 address <MAAS:New> <maas-enlist (Ubuntu):New> <https://launchpad.net/bugs/1717287> | 21:41 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!