id	summary	reporter	owner	description	type	status	priority	milestone	component	version	resolution	keywords	cc	launchpad_bug
1156	the test_2_good_8_hung_then_1_recovers_* methods in test_hung_server are wrong	davidsarah	somebody	"In changeset [changeset:3e4342ecb3625899] (6 months ago at time of writing, but I missed it at the time), a change was made to the [source:src/allmydata/test/test_hung_server.py@4611#L280 test_2_good_8_hung_then_1_recovers_* methods] in {{{test_hung_server.py}}} that defeated the intent of the tests. The intent was to have 8 servers that are initially hung (so that there are not the 3 required shares available), then have one of the servers recover after some period of time, and check that the download continues.

[changeset:3e4342ecb3625899] made the recovery happen immediately, so that the tests are equivalent to the {{{test_3_good_7_hung}}} ones. I'm not sure how useful these tests ever were, but they are definitely not useful with this change. We should either fix or remove them."	defect	new	normal	eventually	code	1.8β		test download