Community
Participate
Working Groups
If 'systemctl restart nftables.service' occurs while eclipse debugging is active (pydev in my case) within a few seconds all the processes under test terminate with a result code 137 = 128 + 9 = signal.SIGKILL I don't know whether it matters but the application under test has a three or so sub-processes, and the systemctl command is run in one of them. But it doesn't matter, even if I run the systemctl restart command manually in another shell entirely the eclipse process under test just stops shortly thereafter having nothing directly to do with nftables. However, if the command is run directly: 'nft -f /etc/nftables.conf' there is no impact on eclipse debugging. Because of the content of the nftables.conf file, both methods flush the ruleset then load it. I suspect it is to do with a weird interaction with the ProtectSystem or ProtectHome settings in the nftable unit files and eclipse operations (the workspace is beneath the home directory tree). Anyway --- trivial to reproduce, 100% recurrence. FWIW, running under kde/ubuntu-eoan/ 5.3.0-29-generic #31-Ubuntu SMP Fri Jan 17 17:27:26 UTC 2020 x86_64 x86_64 x86_64 GNU/Linux Java is openjdk-11-jre Thanks for a great platform! Harry Coin, Bettendorf, Iowa
