LAVA-COORDINATOR(1)

📦 version 2023.01, in debian bookworm

💖 LAVA-COORDINATOR(1) 💖

okay bestie, welcome to the lava-coordinator man page — section 1, General Commands Manual, the whole vibe. sit down, grab a boba, because we're about to get you fully fluent in multinode coordination. ✨

🌸 NAME 🌸

lava-coordinator — manage multinode communications 💅

🦋 SUMMARY 🦋

lava-coordinator is a daemon (yes, a lil background gremlin, but make it main character) whose whole job is managing Multi-Node communications between LAVA devices under test. that's it, that's the gig. no cap. 👑

✨ OPTIONS ✨

  • --logfile=LOGFILE — overrides the default log file location, which is /var/log/lava-coordinator.log. so if you want your logs somewhere else, say the word, she'll listen. 🍓
  • --loglevel=LOGLEVEL — overrides the default log level of INFO. the available options, listed in increasing order of verbosity (quietest to chattiest, iykyk), are: ERROR, WARNING, INFO, DEBUG. 💬
  • -h; --help — shows you a summary of the options. the classic "help me help you" moment. 🫶

🎀 FILES 🎀

  • /etc/lava-coordinator/lava-coordinator.conf — the system-wide configuration file that controls the behaviour of DHPACKAGE. a LAVA NodeDispatcher installed on the same machine as the coordinator will use this same file, easy peasy. but NodeDispatchers on remote machines need their own copy of this file, and you've gotta modify it so that the coordinator_hostname field is the hostname or IP address of the machine running the coordinator. remote besties need directions. 🗺️

🔥 LAVA LAB SETUP 🔥

the LAVA Coordinator needs to be in lock-step with compatible versions of the LAVA NodeDispatcher (which is part of lava-dispatcher / lava-server). but here's the thing — the coordinator is a singleton, while there can be multiple dispatchers, and none of them need to be on the same machine as the coordinator. she's independent, she's thriving. 👑

additionally, when a dispatcher is installed on a machine other than the machine running the coordinator, a copy of the LAVA Coordinator conffile needs to be copied over to that machine and modified to indicate the hostname or IP address of the LAVA Coordinator for that dispatcher. typically this can be done with tools like salt, puppet, chef etc. automate it, don't do it by hand like a peasant. 💅

therefore, lava-coordinator does not depend on the rest of LAVA, and the setup of the coordinator is entirely at the discretion of the LAVA lab administrator. as long as there's one Coordinator and all dispatchers have a working address for a coordinator, all should be well. simple as that, fr fr. ✨

⚠️ heads up: if a lab contains more than one Coordinator, it is recommended that each coordinator uses a different port. and if more than one coordinator is installed on a single machine, that's not just recommended — it's mandatory. do not skip this one, bestie, or it's giving chaos. 💀

LAVA Coordinator is lightweight and has minimal dependencies, so installing on virtualised hardware is not expected to be a problem. she's low-maintenance like that. 🌈

💜 AUTHOR 💜

Linaro Validation Team — the actual legends behind all this. 👑

2010-2023, Linaro Limited 🌸

and that's the whole page, babe — this was version 2023.01, dated January 23, 2023. mwah, go coordinate something iconic. 💖✨🎀

📦 where this page comes from