| === wpk_ is now known as wpk | ||
| Chipaca | hello all | 10:15 |
|---|---|---|
| Chipaca | is it known that the maas snap dumps core on install on 16.04? | 10:15 |
| Chipaca | https://pastebin.ubuntu.com/26415649/ | 10:16 |
| roaksoax | Chipaca: did you install with --devmode ? | 10:19 |
| Chipaca | roaksoax: no | 10:20 |
| roaksoax | Chipaca: https://docs.ubuntu.com/maas/2.3/en/installconfig-snap-install | 10:22 |
| Chipaca | why are we telling people to install stuff with --devmode :-( | 10:22 |
| Chipaca | but, anyway, that answers my question; it's known broken. onwards. | 10:23 |
| roaksoax | Chipaca: because otherwise maas woudldn't work | 11:09 |
| roaksoax | there are blockers that prevent the maas snap (not in maas) from working in fully confined mode | 11:10 |
| roaksoax | so we have to use devmode | 11:10 |
| vogelc | roaksoax: We upgraded our prod to 2.3.0 last night and now we get out of memory errors on the region controller when someone restarts the region controller or does a refresh on the ui. Thoughts? | 13:17 |
| vogelc | roaksoax: builtins.OSError: [Errno 12] Cannot allocate memory | 13:17 |
| roaksoax | vogelc: do you have more details than that ? e.g. regiond.log and process list, etc | 13:21 |
| vogelc | roaksoax: sure, I will get them posted to a paste bin | 13:21 |
| vogelc | roaksoax: https://pastebin.com/sBG42Grk | 13:33 |
| vogelc | roaksoax: We increased the memory on the vm to 24GB and it would still consume all the memory and kill the process. | 13:34 |
| roaksoax | vogelc: does a top or similar show what is utilizing cpu/memory ? | 14:05 |
| vogelc | roaksoax: https://pastebin.com/97cchLQ8 | 14:16 |
| vogelc | roaksoax: that is top and syslog output. I just issued a refresh in the web ui | 14:16 |
| xygnal | roaksoax: would appreciate what actions we can take to debug why vogelc and I are seeing this. It's as if the system is queueing up load until memory runs out. How do we see what 'load' is going on internally? | 15:38 |
| xygnal | roaksoax: listing nodes in webui takes a long time, and seems to stall at 500. never finishes 'loading' status. | 16:25 |
| xygnal | roaksoax: if i refresh the page, shortly thereafter twistd3 consumes all memory and is killed by kernel. | 16:27 |
| xygnal | within minutes | 16:27 |
| bladernr | is there a way to query maas via cli and determine what kernel version is available for each option in maas? (something that does "Show me the kernel verision installed by 16.04-hwe") | 17:42 |
| bladernr | ^^ I'd like to be able to easily determine exactly what version MAAS will install if I choose hwe vs hwe-edge (currently BOTH seem to be 4.13, but they often don't match up) | 18:00 |
| gunix | bladernr: what is the difference between hwe and hwe edge? | 18:07 |
| gunix | i know kvm vms work only with hwe ... but no idea what edge is | 18:07 |
| bladernr | hwe is the current supported HWE kernel, -edge is the NEXT one coming | 18:08 |
| ltrager | bladernr: there isn't a way to do it with MAAS however you could query apt which is where the kernel comes from | 18:43 |
| ltrager | bladernr: e.g apt-cache policy linux-image-generic-hwe-16.04-edge shows its 4.13 | 18:43 |
| bladernr | ltrager, thanks, that'll work. I was hoping for a MAAS Cli way to ask it about the things it's going to install, but that too works | 19:53 |
| mup | Bug #1532286 changed: Postgresql-9.5 breaks MAAS <MAAS:Invalid> <https://launchpad.net/bugs/1532286> | 21:05 |
| klj1218 | I'm trying to create a custom curtin preseed file and I've noticed that if I include "system_id" for the {node} it fails to pick the file up... I have to use the "hostname" instead. This doesn't match the documentation. | 21:38 |
| klj1218 | fail --> curtin_userdata_centos_amd64_generic_centos70_q6g3hd | 21:39 |
| klj1218 | work --> curtin_userdata_centos_amd64_generic_centos70_dell01-srv01 | 21:39 |
| klj1218 | The template naming doc is here: https://docs.ubuntu.com/maas/devel/en/nodes-custom | 21:41 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!