| lubot | <HMollerCl> iamges are gone | 00:56 |
|---|---|---|
| lubot | <HMollerCl> So, I can easily parse tasks, but I don't have the time to investigate how the api works | 00:56 |
| lubot | <HMollerCl> I know there can be loaded as json | 00:57 |
| lubot | <HMollerCl> likde this https://secure.phabricator.com/book/phabricator/article/conduit_edit/ | 00:59 |
| lubot | <HMollerCl> but how to connect, to phab to load this files is something I don't have the time to do it. | 00:59 |
| === meetingology` is now known as meetingology | ||
| === ddevault_ is now known as ddevault | ||
| lubot | <teward001> @HMollerCl wxl: As Phab is in an LXD container once we're 'satisfied' with the configuration a weekly 'copy' job is going to run that will copy the *entire* Phab instance's container from the system it's now on over to three LXD instances I mantain as 'backups' - this way we have full data copies rather than having to rely on the ph | 14:52 |
| lubot | abricator specific toolsets, we'll just ahve the 'full system containers' copied over. | 14:52 |
| lubot | <teward001> that way instead of having to redeploy Phab and then redeploy the data to it, we just copy the container over, turn it on, and "everything works" again | 14:52 |
| lubot | <teward001> it's an LXD approach to backups :P | 14:53 |
| lubot | <teward001> but so far has worked for my LXD clusters at large (15+ containers and counting!) | 14:53 |
| lubot | <teward001> this also works if we want to spin a second dedicated 'backup' system as well | 14:54 |
| lubot | <teward001> because we can just `lxc copy lubuntu-main:... local:` to copy from the 'main' server back down to a backup server, and vice versa. | 14:54 |
| lubot | <HMollerCl> What's lxd? | 15:03 |
| lubot | <acheronuk> @HMollerCl [What's lxd?], https://linuxcontainers.org/lxd/introduction/ | 15:08 |
| lubot | <teward001> ^ we're using lxd containers :P | 15:09 |
| lubot | <teward001> which was Simon's suggestion :P | 15:09 |
| lubot | <HMollerCl> Aaa, OK thanks! | 15:10 |
| lubot | <teward001> so while the 'host' system has a single IP address | 15:10 |
| lubot | <teward001> that system has an nginx instance that actually handles the SSL certs, etc. | 15:10 |
| lubot | <teward001> and hands off to the 'backend' containers for phab, etc. | 15:10 |
| lubot | <teward001> but each actual *service* sits independent of each other within their own containers so as not to 'pollute' the host system's libraries, etc. | 15:11 |
| lubot | <teward001> but that also means we can independently move thes ervices to other servers without having to do a from-scratch deployment :P | 15:11 |
| lubot | <teward001> I DO have to ask stgraber how to live-migrate though, we've got a couple of headaches with regards to LXD and the snap where live migration doesn't work anymore... | 15:11 |
| lubot | alex linux was added by: alex linux | 22:04 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!