Neofs-cli control command hangs up if called immediately after storage node start

[vdomnich-yadro]

Minor issue, has something to do with race condition on startup.
We did workaround this in tests by adding a small delay between node start and calling neofs-cli command.

Steps to Reproduce (for bugs)

Given dev-env environment with all 4 storage nodes running and online.

echo 'password: ""' > config.yaml

neofs-cli --endpoint s02.neofs.devenv:8081 -w services/storage/wallet02.json -c config.yaml control set-status --status offline

# Tick epoch
sleep 10s
bin/tick.sh

# Remove all data from storage volume
docker stop s02
sleep 5s
rm -rf /var/lib/docker/volumes/storage_storage_s02/_data/*

# Start container using python (it seems to be faster than docker start command and is necessary to reproduce race condition)
python3 -c "import docker; d = docker.APIClient(); d.start('s02');"
sleep 0.4s  # 0.4s was selected experimentally, it happens after grpc api is up, but before morph notifications start working

# This command hangs up:
neofs-cli --endpoint s02.neofs.devenv:8081 -w neofs-dev-env/services/storage/wallet02.json -c config.yaml control set-status --status online

Current Behavior

Once set-status command from the script times out, all subsequent calls of neofs-cli control are failing as well. The docker container needs to be restarted in order to bring things into normal state.

Expected Behavior

The set-status command from the script above does not hang up. Or, at least, it does not leave node in the broken state and subsequent calls to set-status are working.

Regression

No

Your Environment

• Version used:
  v0.31.0
• Server setup and configuration:
  Local devenv
• Operating System and version (uname -a):
  Linux ubuntu 4.15.0-189-generic #200-Ubuntu SMP Wed Jun 22 19:53:37 UTC 2022 x86_64 x86_64 x86_64 GNU/Linux

[vdomnich-yadro]

If we increase delay between python code that starts container and neofs-cli control command, then everything works.
Attached are logs for hang-up case (log-bad.txt) and for case with increased delay, when everything works (log-good.txt).

[roman-khimov]

Should be OK now with reordered init, #2585.