INDY-1911: add integration tests #1034
Conversation
Signed-off-by: toktar <renata.toktar@dsr-corporation.com>
Signed-off-by: toktar <renata.toktar@dsr-corporation.com>
|
test this please |
| ensure_all_nodes_have_same_data(looper, nodes=txnPoolNodeSet) | ||
|
|
||
| for node in txnPoolNodeSet: | ||
| assert node.viewNo == view_no |
There was a problem hiding this comment.
Since we are already relying in this test on a fact that view change starts on some nodes I think it's better to do something like this:
| assert node.viewNo == view_no | |
| assert node.viewNo > view_no |
Another option is to just get rid of this check, since actually all that we want here is to check that nodes will eventually come to consensus after some panics and restarts, and this is already done by ensureElectionsDone and ensure_all_nodes_have_same_data.
| nodes_to_restart = txnPoolNodeSet[1:3] | ||
|
|
||
| # waiting to discard InstanceChange | ||
| looper.runFor(tconf.OUTDATED_INSTANCE_CHANGES_CHECK_INTERVAL) |
There was a problem hiding this comment.
Probably it's better to wait for a little bit more than timeout interval:
| looper.runFor(tconf.OUTDATED_INSTANCE_CHANGES_CHECK_INTERVAL) | |
| looper.runFor(tconf.OUTDATED_INSTANCE_CHANGES_CHECK_INTERVAL +1) |
Another option (although a bit more intrusive) is to actually wait until nodes don't have INSTANCE_CHANGE from panic_node anymore.
There was a problem hiding this comment.
Change to wait INSTANCE_CHANGE removing
| def check(): | ||
| assert all(panic_node.view_changer.instanceChanges.has_inst_chng_from(view_no + 1, node.name) | ||
| for node in nodes_to_restart) | ||
| assert not panic_node.view_changer.instanceChanges.has_inst_chng_from(view_no + 1, panic_node.name) |
There was a problem hiding this comment.
Second assert looks a bit excessive here. Also in first assert it might make sense to actually check that all nodes received instance changes from nodes_to_restart, not just panic_node
There was a problem hiding this comment.
In the second assert we check that panic_node hasn't InstanceChange from itself.
There was a problem hiding this comment.
Remove this assert because added this check before nodes restarting
| looper.removeProdable(name=node_to_disconnect.name) | ||
|
|
||
| # add node_to_disconnect to pool | ||
| node_to_disconnect = start_stopped_node(node_to_disconnect, looper, tconf, |
There was a problem hiding this comment.
Are we sure that this is enough to clear state of node_to_disconnect?
There was a problem hiding this comment.
Yes, beacause start_stopped_node create a new TestNode
Signed-off-by: toktar <renata.toktar@dsr-corporation.com>
Signed-off-by: toktar <renata.toktar@dsr-corporation.com>
Signed-off-by: toktar <renata.toktar@dsr-corporation.com>
Signed-off-by: toktar renata.toktar@dsr-corporation.com