This will also have the benefit of getting fresh metadata only once per transition for a given resource agent. Currently, the agent meta-data command is invoked for every resource start, even if multiple resources share the same agent.

This would be worth backporting to 2.1 if easy

As an example, for PCMK_panic_action, this will involve creating an enum for the value, which will minimize the work that panic_local() has to do. (The "sync" portion of it should likely be a flag or bool separate from the enum.)

Currently lrmd_remote_client_msg() calls pcmk__read_remote_message(), which loops over read_available_remote_data() up to the timeout.

In T855#13286, @clumens wrote:

My first thought here is to start by converting pcmk__remote_ready to be an async function that checks the remote to see if it's ready once and sends a result up to the mainloop. However, one caller of this function is lrmd_poll, which we do not use anywhere but I assume is public API. Do we need to continue keeping it (and lrmd_dispatch) around? If so, we might need sync and async versions of this function.

In T837#13290, @clumens wrote:

g_log_set_always_fatal((GLogLevelFlags) 0); /*value out of range */