PSI/e2e infra updates #570
Conversation
This is useful for adding more instances after already deployed ones.
Fedora 39's moreutils-parallel has a completely different CLI now. It
gets the `{}` values as arguments after `--` instead of reading them
from stdin.
shared_net_7 has some intermittent undiagnosed networking trouble, see Red Hat SNOW ticket 5c1149998723359ca92b4156cebb35a0
ansible/psi/deploy-all.sh
Outdated
|
|
||
| ansible-playbook -i inventory/ psi/image-cache.yml | ||
| ansible-playbook -f20 -i inventory/ psi/image-cache.yml |
There was a problem hiding this comment.
non-blocking nit, --forks would be what I prefer reading here. 20 sounds like a lot then again never seen ansible-playbook runs "eat resources" with the default 5.
There was a problem hiding this comment.
On the client side it's really just a bunch of ssh processes. I find 20 to be just fine for all our cockpituous cases, we don't do anything expensive on the caller side.
But I'll change to --forks indeed.
I did this manually in the previous install, and forgot to automate it.
|
@mvollmer commits 1, 2, 4 should be obvious; 3 is the broken networking that I reported to the admins, and so far changing the network to |
This got rolled out already.