QA-607: Fix for wait_for_upgrade_cluster test #542
Conversation
| for table_row_num in [1, 2, 3]: | ||
| row = {} | ||
| table.append(row) | ||
| for table_column in range(1, cluster_nodes): | ||
| for table_column in range(1, 6): # for table_column in range(1, cluster_nodes): |
There was a problem hiding this comment.
can we go back to cluster_nodes so we have a single source of truth for this?
There was a problem hiding this comment.
the very reason I made this change is that passed cluster_nodes has incorrect value - 3 - which in turn leads to incomplete data in the table (only consisting of row["name"], row["url"])
cluster_nodes number is taken from self.selenium_runner.props.cluster_nodes - I didn't really have time to debug this
There was a problem hiding this comment.
it can be calculated from
count = 0
for starter in self.starter_instances:
count += len(starter.expect_instances)
agents are counted here as well?
There was a problem hiding this comment.
I will change the number of nodes in the test case - so that the method of the (Node) page class itself is not impacted (and can be safely called from other tests)
There was a problem hiding this comment.
I also don't understand why we have range(1, cluster_nodes) there in the first place - it is plain wrong as we're always supposed have 5 columns and 5 data 'points' in the table:
[row["name"], row["url"], row["version"], row["date"], row["state"]]
I guess cluster_nodes number was meant to be used in the outer loop (loops over the number of rows/nodes) :
for table_row_num in [1, 2, 3]:
This PR contains code changes necessary to fix
test cases to check the integrity of the old system after the upgrade (Cluster)test case that has been failing for a long while.UI tests
arangodb-devel-UI.Linuxjob succeeded - build 428