qa/vstart_runner: fix mutiple 'timeout' option conflicting
The 'timeout' option in the environment may will conflict with the
ones in some ceph commands, like:
$ timeout 120 ./bin/ceph daemon mds.b session config 8718 timeout 45
And the old code will also give us incorrect result like:
['adjust-ulimits', 'ceph-coverage', 'timeout', '120', 'ceph', 'fs', 'dump']
will be transfered to:
['adjust-ulimits', 'ceph-coverage', '120', 'ceph', 'fs', 'dump']
The '120' is left behind.
Fixes: https://tracker.ceph.com/issues/44437
Signed-off-by: Xiubo Li <xiubli@redhat.com>
(cherry picked from commit
1bcd319496f507f76864a972c1de63439dd78d4a)