The test failed because the linux user it used was not available
on buildbot. Now the test is part of the larger pam_authentication-
test, which sets up the users.
As Xpand does not have a mysql.user table but a system.users
table, we need to arrange things so that the query used for
selecting annymous users can be different depending on the cluster.
Needs to be a member function as the dropping of anonymous
users must be done differently for Xpand. With current code,
no anonymous users in Xpand will be dropped, and there will
always be an error (so, should someone care about the return
code, this would not work).
Xpand nodes must be added one by one to the cluster as an attempt
to add more nodes will fail, if any of the nodes to be added
already are in the cluster.
Further, the adding of a node may also fail if, when the addition
is made, the addition of the previous node is still in process.
Now it will be attempted at most 5 times, while sleaping as many
seconds between each attempt as there has been attempts.
Some additional logging was also added.
The assumption that errno would be the same for the duration of the whole
call chain is wrong as it's possible that other calls that modify it are
made. Using the actual error string itself is much more reliable.
Every change to MariaDB Enterprise should be tested with Maxscale.
Maxscale should be stable - installed from production repository.
New enviromental variable 'maxscale_product' is introduced.
It can be 'maxscale' or 'maxscale_ci' (default is 'maxscale_ci')
config_test expects Maxscale start failure, but ssh to
Maxscale node also fails if VM is broken.
Test should fail in case of VM problem to prevent
Maxscale problem hiding.
