CASSANDRA-21552 Force repair should ignore min_repair_interval - #5004
CASSANDRA-21552 Force repair should ignore min_repair_interval#5004tommystendahl wants to merge 6 commits into
Conversation
99c770d to
d4c3d53
Compare
|
I created a new test case for end-to-end using |
| @@ -196,7 +196,8 @@ public void repair(AutoRepairConfig.RepairType repairType) | |||
| UUID myId = Gossiper.instance.getHostId(FBUtilities.getBroadcastAddressAndPort()); | |||
|
|
|||
| // If it's too soon to run repair, don't bother checking if it's our turn. | |||
There was a problem hiding this comment.
@tommystendahl could you please move this comment where it should be? If you read how it is now, that comment seems to be out of place because we changed the implementation but the comment stayed and it is confusing.
There was a problem hiding this comment.
When I did it it made sense to me to leave the comment as it was but looking again I can see how it could be confusing but I don't think just moving it will help. I added an new comment for tooSoonToRunRepair in the same way as I did for shouldSkipRepairDueToInterval and I rewrote this comment.
| assertTrue("repair_finish_ts should advance after force repair runs, but was " | ||
| + finishTimeBefore + " -> " + finishTimeAfter, | ||
| finishTimeAfter > finishTimeBefore); | ||
|
|
There was a problem hiding this comment.
Pleae verify that
assertFalse(AutoRepair.instance.shouldSkipRepairDueToInterval(repairType, repairState, config, myId));
|
|
||
| // Record the finish time before repair runs | ||
| long finishTimeBefore = AutoRepairUtils.getLastRepairTimeForNode(repairType, myId); | ||
|
|
There was a problem hiding this comment.
Pleae verify that
assertTrue(AutoRepair.instance.shouldSkipRepairDueToInterval(repairType, repairState, config, myId));
patch by Tommy Stendhal; reviewed by for CASSANDRA-21552