search

oar-team / oar3

OAR: versatile resource and job manager for cluster (third generation)
Other
8 stars 11 forks source link

Systemctl stop is not stopping all the processes #56

Closed bzizou closed 3 weeks ago

bzizou commented 2 months ago

When stopping oar3 with systemctl stop oar-server, greta processes remain and cause the server to fail on the next systemctl start oar-server with the following error:

[   ERROR] [2024-07-12 09:59:25,020] [oar.modules.greta::__init__:158]: Failed to bind Greta endpoint: Address already in use

We also noticed a bipbip and an appendice process, also remaining after a stop:

oar      1184382  0.0  2.7 151364 55780 ?        Sl   10:21   0:00 /usr/bin/python3 /usr/lib/oar/oar-appendice-proxy
oar      1184383  0.0  2.8 152868 58120 ?        Sl   10:21   0:00 /usr/bin/python3 /usr/lib/oar/oar-bipbip-commander

So we have to manually kill those processes (oar-greta, oar-appendice-proxy and oar-bipbip-commander) to successfully restart the oar-server after a stop.

bzizou commented 3 weeks ago

Fixed by ea5c57b7719f5fbed89682e63d2c86a4b23bdb9c and 1872491fafdaeafc6a4fcec0e31cc0b545a80d28 and debian/3.0 branch