Is it unusual to see two nailgun messages ?

❯ pants package ::
19:16:51.52 [INFO] (pe_nailgun::nailgun_pool) Initializing Nailgun pool for 40 processes...
19:16:54.63 [INFO] (pe_nailgun::nailgun_pool) Initializing Nailgun pool for 40 processes...

each time we run pants ?

I'm not directly familiar with Nailgun, but this seems to make sense. I see the same behavior with eg. flake8, isort, black, etc. It just means pants is chunking up your input targets for the sake of parallelization.

a pair of (40) parallelisation backends 😄 ?

This is different: it’s not batching of linters but a single preliminary step (that only happens once per pantsd restart). And yes, I always see two of these as well, not sure why.

So perhaps a related topic then. pantsd .. always restarts. that is .. it is never remaining running in the background; e.g. and pants always takes ~10s to bootstrap.

~/pro/git/repo on  mybranch [$!?]
❯  time pants version
08:41:36.03 [INFO] (pe_nailgun::nailgun_pool) Initializing Nailgun pool for 40 processes...
08:41:39.31 [INFO] (pe_nailgun::nailgun_pool) Initializing Nailgun pool for 40 processes...
2.22.0
pants version  9.51s user 0.57s system 100% cpu 10.065 total

~/pro/git/repo on  mybranch [$!?] took 10s
❯  time pants version
08:41:47.21 [INFO] (pe_nailgun::nailgun_pool) Initializing Nailgun pool for 40 processes...
08:41:50.48 [INFO] (pe_nailgun::nailgun_pool) Initializing Nailgun pool for 40 processes...
2.22.0
pants version  9.72s user 0.56s system 100% cpu 10.270 total

• Should I expect the pantsd to remain ? • I presume that 10s is very poor ; (I should investigate)

Ah yes, if pantsd is restarting every run that will kill your performance