improve the SAGE_ROOT/local/bin/sage-cleaner script so that it kills those damn lisp.run processes

[williamstein]

Fact is, several people have observed lisp.run's getting left running.

If anybody can find a 100% reliable way to replicate this problem, PLEASE LET US KNOW!

CC: @williamstein @mwhansen @robertwb

Component: misc

Issue created by migration from https://trac.sagemath.org/ticket/2518

[sagetrac-mabshoff]

comment:1

The issue here is likely that we use

os.kill(int(pid),0)

It is likely that using "kill -9 $PID" via os.system() ought to fix the issue.

Cheers,

Michael

[sagetrac-mabshoff]

comment:2

I believe the problem was that if clisp crashed and restarted those processes could get stuck at 99% and not be killed at exit.

This hasn't been reported in a while, so I would prefer to close this as fixed. Thoughts?

Cheers,

Michael

[TimDumol]

comment:3

Since this hasn't been reported in around 8 months, I think this should be closed.

[mwhansen]

comment:4

Seems reasonable to me.