#Work distribution/ Remote execution

1 messages · Page 1 of 1 (latest)

rancid warren
#

Just found out about dagger.io, so I'm trying to figure out what I can and cannot do, and if it'll fit my needs.

I've been looking for a tool like this, where you can run pipeline or builds both locally and in CI. dagger.io does seem to do a fairly good job. But I am unable to find good information about its parallelism, work distribution and remote execution.

Basically (TL;DR), can dagger.io distribute tasks on remote machines? Where can I find docs about it?
Because building all software on the same machine/host would take too much compute power/time.

novel temple
# rancid warren Just found out about dagger.io, so I'm trying to figure out what I can and canno...

Hey there! Dagger doesn't currently replace your end-to-end CI stack. When adopting Dagger (https://docs.dagger.io/adopting), you still need some sort of compute platform (generally your current CI infrastructure) to run your Dagger pipelines. Dagger does a really good job on de-dupping, caching and parallelizing your steps on the environment that you run it but if you need to distribute your work across multiple machines, the recommended approach is levarage your current CI platform to achieve that.

rancid warren
#

That's a bit of a shame. I was hoping dagger.io would be an alternative to bazel. Which has both caching and remote execution to be able to build large softwares (locally) and in CI.

The problem I see with using the CI platform to distribute the workload is that it's a pain to build the software locally then. You will be forced to push your code and have the CI build it. Which I thought was the problem dagger.io was trying to solve.

Perhaps only an issue with large and computationally heavy SW builds.