Fork me on GitHub
#onyx
<
2018-06-21
>
joelsanchez20:06:30

how does onyx manage to release everything with consistent versions? I have just 3 artifacts and it's hell already

lucasbradstreet20:06:36

Investment in build tooling

lucasbradstreet20:06:35

Everything is done through CI and scripts. Main onyx release will also update the plugins, run tests and release those too

😮 4
lucasbradstreet20:06:00

We also have scripts in the plugins that let us release updates independent of an onyx release, so we can update those too

lucasbradstreet20:06:37

You can also checkout the circle.yml to see what it does on CI after that script runs

lucasbradstreet20:06:41

it was a fair bit of work getting there but it paid off for us enormously over the long run.

lucasbradstreet20:06:54

If I started again and used a mono repo I’d probably use https://github.com/amperity/lein-monolith

joelsanchez20:06:20

I also found this which clears up much of the mystery too https://github.com/onyx-platform/onyx-release-scripts

joelsanchez20:06:53

thanks for the links, will try lein-monolith - it may be good enough for me 🙂

michaeldrogalis21:06:32

If I could do it again, I'd have gone with a monorepo.

michaeldrogalis21:06:53

I wanted separate projects in the beginning to separate issue trackers, but that never ended up being worthwhile.

lucasbradstreet21:06:06

It worked out fine in the end, but it was a lot more work.

michaeldrogalis21:06:28

deps.edn supplants a lot of the manual work we're doing with lein-voom.

sreekanth22:06:13

Hi, when am trying run latest dashboard with onyx 0.12.2.0 org.apache.zookeeper.KeeperException$NoNodeException: KeeperErrorCode = NoNode for /onyx/<tenancy-id>/job-scheduler/scheduler code: -101 path: “/onyx/eone-cp-flow/<tenancy-id>/scheduler”

sreekanth22:06:15

any help on this?

lucasbradstreet22:06:42

Did you switch out that tenancy ID in the error?

lucasbradstreet22:06:09

K. Looks like it can’t find the data in zookeeper for that tenancy

lucasbradstreet22:06:25

If your peers never managed to start up properly then they won’t have had a chance to write that data out

sreekanth23:06:42

thanks, i will look into it.