On Thu, 16 Apr 2020 at 11:33, Brad Hubbard <bhubbard@redhat.com> wrote:
couldHi Rishabh,
I tried to reproduce this on SHA d65516a020720726d6e6afef6fdfde699890f719 but did not see the issue you report.
I ran the following.
$ ./do_cmake.sh -DWITH_PYTHON2=OFF -DWITH_CEPHFS_SHELL=ON -DWITH_BABELTRACE=OFF -DWITH_MANPAGE=OFF -DWITH_RBD=OFF -DWITH_RADOSGW=OFF -DWITH_KRBD=OFF $ cd build $ make -j8 vstart-base $ MON=1 MDS=0 MGR=1 OSD=3 ../src/vstart.sh -n -x -l -d --without-dashboard
Everything worked as expected. I then stopped and started vstart three times as your OP implied that vstart seems to work at least once when you build from scratch? Perhaps you could describe the events between it working and not working or suggest other ways I might be able to reproduce the issue?
Most of the stuff I did was pretty usual: modified a few Python and C++ files, ran "make -j 7" and tested my changes using vstart_runner.py.