=== narindergupta is now known as narinderguptamac | ||
TimM[m] | Hi all, would anyone be interested in collaborating to charm up Jitsi and BigBlueButton? | 00:32 |
---|---|---|
thumper | wallyworld: ping | 02:29 |
wallyworld | yo | 02:29 |
thumper | I think you have a bug in the 2.8 upgrade step for the tasks sequence | 02:30 |
thumper | shouldn't the tasks sequence be bumped for every model? | 02:30 |
wallyworld | ok, will fix, what's the issue? | 02:30 |
wallyworld | probs, i'll need to check | 02:30 |
thumper | do we create the task id from the sequence value +1 ? | 02:31 |
thumper | is that why we're bumping the sequence id? | 02:31 |
wallyworld | the sequence starting value and increment logic changed | 02:32 |
wallyworld | so to avoid doubling up when using 2.8 on a 2.7 db, the sequence was incremented | 02:32 |
wallyworld | otherwise the latest sequence number could have been used twice | 02:32 |
thumper | that didn't answer my question | 02:32 |
thumper | it doesn't make sense to me why we actually have this upgrade step | 02:33 |
wallyworld | because we went from calling sequence() to sequenceWithMin() | 02:34 |
wallyworld | so starting id is 1 not 0 | 02:35 |
wallyworld | and if you compare the old logic vs new logic, running 2.8 would have reused an existing task id | 02:36 |
wallyworld | i haven't got the exact code handy | 02:36 |
* thumper is looking at it now | 02:37 | |
wallyworld | the code has moved as well | 02:37 |
wallyworld | rather, additonal use of task sequence added | 02:38 |
thumper | ok | 02:39 |
wallyworld | action id used to add 1 to sequence() | 02:39 |
thumper | I do think that the upgrade step needs to be run across all models | 02:39 |
wallyworld | which started from 0 | 02:39 |
thumper | right now it only does the controller model | 02:39 |
wallyworld | yeah, it does | 02:39 |
wallyworld | which is a bug :-( | 02:39 |
* thumper is writing an upgrade step now for unit machine id denormalisation | 02:39 | |
wallyworld | drive by? :-) | 02:40 |
TimM[m] | 1st pass at a "juju personas" document https://discourse.jujucharms.com/t/juju-user-personas/2808 | 03:50 |
jam | manadart: ping | 07:56 |
jam | or maybe stickupkid | 07:57 |
=== balloons6 is now known as balloons | ||
sdhd-sascha | Hi, i use juju with an lxd cluster. Now i remove one lxd-server from the cluster. Where can i find the configuration for the ip address ? | 10:24 |
sdhd-sascha | I searched and changed all occurrences at `~/.local/share/juju` | 10:24 |
sdhd-sascha | But i still get this error, when i try to remove or add unit's. Or when i try to call `juju storage` | 10:24 |
sdhd-sascha | `ERROR getting state: getting storage provider registry: Get https://...:8443/1.0: Unable to connect to: ...:8443` | 10:24 |
stickupkid | sdhd-sascha: is this the same one you signed up to? | 10:26 |
sdhd-sascha | stickupkid: it was the initial lxd-server from the cluster, where i later bootstrapped juju. | 10:56 |
stickupkid | sdhd-sascha, yeah, so I believe the issue is that we only know about that ip. To change the ip of that will require so mongo surgery. I think it's not unreasonable to make a bug for this. | 10:58 |
stickupkid | manadart, achilleasa_ that's correct ^ | 10:59 |
stickupkid | ? | 10:59 |
sdhd-sascha | stickupkid: where can i find the mongodb? | 10:59 |
manadart | stickupkid: Yep. | 10:59 |
stickupkid | sdhd-sascha, https://discourse.jujucharms.com/t/login-into-mongodb/309 | 11:00 |
sdhd-sascha | stickupkid: super. Thank you :-) | 11:00 |
stickupkid | sdhd-sascha, would you mind taking the time to file a bug? https://bugs.launchpad.net/juju/+bugs | 11:00 |
sdhd-sascha | stickupkid: yes, i will. Maybe i have time to create a patch, too | 11:01 |
stickupkid | sdhd-sascha, I'm unsure what's the best way to solve this, maybe the lxd provider/environ should help the instancepoller | 11:02 |
sdhd-sascha | Ah, ok. I will see | 11:02 |
sdhd-sascha | Thank you | 11:03 |
stickupkid | manadart, jam you seen this stack trace before ? https://paste.ubuntu.com/p/djJKY7dszN/ | 11:39 |
jam | 2020-03-18 11:38:45 INFO juju.cmd supercommand.go:83 running jujud [2.7.5.1 9825e246a9ec70e6551744d033802f19d78cabce gc go1.14] | 11:40 |
jam | runtime: mlock of signal stack failed: 12 | 11:40 |
jam | runtime: increase the mlock limit (ulimit -l) or | 11:40 |
jam | runtime: update your kernel to 5.3.15+, 5.4.2+, or 5.5+ | 11:40 |
jam | fatal error: mlock failed | 11:40 |
jam | stickupkid: ^ I have not seen that before. | 11:40 |
jam | sounds like a bug in focal if I had to guess | 11:40 |
stickupkid | jam, neither have i | 11:40 |
manadart | stickupkid: Nope. | 11:40 |
stickupkid | jam, just wrapping up my focal PR, I'll see if it happens again | 11:41 |
stickupkid | manadart, CR my changes since thumper last reviewed - https://github.com/juju/juju/pull/11332 | 12:18 |
stickupkid | manadart, also I need help testing if possible | 12:18 |
manadart | stickupkid: Yep. Gimme a few. | 12:19 |
hml | stickupkid: 1 line review pls? https://github.com/CanonicalLtd/juju-qa-jenkins/pull/402 | 12:22 |
stickupkid | manadart, whilst you're there - https://github.com/juju/juju/pull/11333/files#diff-e8543713fc0c30ea33131d41edf815e7R16 | 13:00 |
manadart | rick_h_: Were we going to hang on and talk test? | 13:33 |
rick_h_ | manadart: yep my bad | 13:35 |
* rick_h_ got distracted | 13:35 | |
hml | anyone noticed something wrong with changes to “juju model-config logging-config” | 15:42 |
hml | i updated logging-config value and it didn’t take effect in my model, though shown in the model config. | 15:43 |
hml | had to bounce jujud for machine and unit agents | 15:43 |
hml | in develop | 15:44 |
danboid | Where is the juju user data stored eg credential/login stuff? | 15:58 |
danboid | or in other words, how do I move a juju account? | 15:59 |
rick_h_ | danboid: so it's in .local/share/juju | 15:59 |
rick_h_ | danboid: to move an accout the best thing is to just juju login to the controller from a new location | 16:00 |
rick_h_ | danboid: but if you need credntials/etc you need the .local/share/juju/credentials.yaml and clouds.yaml | 16:00 |
danboid | rick_h_, Thanks | 16:01 |
danboid | rick_h_, I presume a single juju user account be used from multiple machines then? | 16:12 |
rick_h_ | danboid: yes, normally if you create a juju user and give them as passowrd you can then juju login with that user/pass as long as you know the IP of the controller | 16:39 |
rick_h_ | danboid: so it's pretty easy to have multiple clients against a single controller | 16:39 |
danboid | Great, thanks | 16:39 |
stickupkid | rick_h_, my PR landed in 2.7 | 17:54 |
stickupkid | rick_h_, and so has manadart's | 17:54 |
rick_h_ | stickupkid: woot woot, will watch for a ci run with that sha then ty! | 17:57 |
stickupkid | nps | 17:57 |
sdhd-sascha | hey, hi, | 19:51 |
sdhd-sascha | i tried to connect to the mongodb, but cannot find the correct password. Like, described here: | 19:51 |
sdhd-sascha | https://discourse.jujucharms.com/t/login-into-mongodb/309/5?u=sdhd | 19:51 |
sdhd-sascha | Hmm, i just inside an juju controller, but can't find the `init.d` or `systemd` startup of the `mongod` ? Who launches this daemon? | 20:00 |
rick_h_ | sdhd-sascha: those are started via systemd as long as you're on ubuntu >= xenial | 20:01 |
sdhd-sascha | rick_h_: thank you. I'm on 20.04 on the host. And the container is `bionic` | 20:02 |
rick_h_ | sdhd-sascha: this is juju bootstrapped on localhost? | 20:02 |
sdhd-sascha | yes | 20:02 |
rick_h_ | sdhd-sascha: so to access the contrller you do "juju switch controller; juju ssh 0" | 20:03 |
rick_h_ | sdhd-sascha: that puts you on the controller machine which has mongodb and jujud running | 20:03 |
sdhd-sascha | inside the controller, this command gives no output ... `# systemctl | grep mongo` | 20:03 |
sdhd-sascha | But, `pstree -p` shows that mongod is running... | 20:03 |
sdhd-sascha | Nope, if i try `juju ssh`, i have this: | 20:03 |
sdhd-sascha | ``` | 20:03 |
sdhd-sascha | $ juju ssh 0 | 20:03 |
sdhd-sascha | ERROR opening environment: Get https://10.0.0.8:8443/1.0: Unable to connect to: 10.0.0.8:8443 | 20:03 |
rick_h_ | sdhd-sascha: what version of Juju? | 20:04 |
sdhd-sascha | rick_h_: that's because i delete `10.0.0.8` before from the cluster... | 20:04 |
sdhd-sascha | wait... | 20:04 |
sdhd-sascha | 3.22 (13840) on every server | 20:04 |
timClicks | I believe the service is called juju-db | 20:04 |
rick_h_ | sdhd-sascha: what is the Juju version though? what does the version in juju status show you? | 20:05 |
rick_h_ | timClicks: +1 | 20:05 |
sdhd-sascha | sorry | 20:05 |
sdhd-sascha | ``` | 20:05 |
sdhd-sascha | juju 2.7.4 10906 latest/stable canonical✓ classic | 20:05 |
sdhd-sascha | lxd 3.22 13840 latest/stable canonical✓ - | 20:05 |
sdhd-sascha | ``` | 20:05 |
rick_h_ | sdhd-sascha: is your contrller up and running? can you do juju status successfully? | 20:06 |
sdhd-sascha | rick_h_: yes | 20:06 |
sdhd-sascha | ``` | 20:06 |
sdhd-sascha | | juju-b1a552-0 | RUNNING | 10.0.2.92 (eth0) | | CONTAINER | 0 | mars | | 20:06 |
sdhd-sascha | +----------------+---------+------------------------+------+-----------+-----------+----------+ | 20:06 |
sdhd-sascha | | juju-b1a552-1 | RUNNING | 10.0.2.94 (eth0) | | CONTAINER | 0 | merkur | | 20:06 |
sdhd-sascha | +----------------+---------+------------------------+------+-----------+-----------+----------+ | 20:06 |
sdhd-sascha | | juju-b1a552-2 | RUNNING | 10.0.2.93 (eth0) | | CONTAINER | 0 | mars | | 20:06 |
sdhd-sascha | ``` | 20:06 |
rick_h_ | sdhd-sascha: is it something you can pastebin? https://paste.ubuntu.com/ | 20:06 |
rick_h_ | sdhd-sascha: so that's lxc list which is showing you the containers you have running | 20:07 |
rick_h_ | looks like you've got a 3 machine HA cluster going? | 20:07 |
sdhd-sascha | rick_h_: what exactly, should i print ? | 20:07 |
sdhd-sascha | rick_h_: yes, 3 HA | 20:07 |
rick_h_ | sdhd-sascha: `juju status` | 20:07 |
sdhd-sascha | https://www.irccloud.com/pastebin/qGvYj9Bz/ | 20:08 |
rick_h_ | sdhd-sascha: hmm, ok can you try that ssh command again with --debug `juju ssh 0 --debug`? | 20:09 |
sdhd-sascha | https://www.irccloud.com/pastebin/ahwq5OaF/ | 20:09 |
rick_h_ | sdhd-sascha: does that make any sense to you? It's confusing to me as the lxd machines are all 10.0.2.xx and not sure what a tomcat port has to do with anything? | 20:10 |
sdhd-sascha | rick_h_: today, your colleage, say's i should send a bugreport about the failure with the deleted lxd-server. | 20:11 |
sdhd-sascha | I will do tomorrow. | 20:11 |
sdhd-sascha | But now, i only want to access the mongodb. But didn't have a password... | 20:11 |
rick_h_ | sdhd-sascha: ok, the services are run via /etc/systemd/system/juju-db.service and /etc/systemd/system/jujud-machine-0.service | 20:12 |
sdhd-sascha | rick_h_: I bootstrapped the lxd-cluster from 10.0.0.8. Then i bootstrapped juju from the same ip. Then i figured out, that the machine has not enough RAM. So i deleted it from the cluster... ;-) ... | 20:12 |
sdhd-sascha | rick_h_: Ah, thank you... i searched for `mongo` | 20:13 |
sdhd-sascha | :-) | 20:13 |
rick_h_ | sdhd-sascha: the db password will be on the machine in the /var/lib/juju/agents/machine-0/agent.conf | 20:13 |
rick_h_ | sdhd-sascha: that post is the right one: https://discourse.jujucharms.com/t/login-into-mongodb/309 but since your juju ssh is wonky you'll have to pull the script apart manually | 20:14 |
rick_h_ | sdhd-sascha: you can skip the juju ssh to the machine since it's lxd and just lxd exec bash on the right instance | 20:14 |
sdhd-sascha | rick_h_: my agent.conf's didn't have a value `statepassword` like i said before | 20:14 |
rick_h_ | juju-b1a552-0 from the status output | 20:14 |
rick_h_ | sdhd-sascha: it should be there. | 20:15 |
sdhd-sascha | rick_h_: ah, thank you :-) i found it ;-) | 20:15 |
rick_h_ | sdhd-sascha: coolio | 20:15 |
sdhd-sascha | :-) | 20:15 |
sdhd-sascha | login works. great +1 | 20:17 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!