1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
|
/**
* Tests that DBClientRS performs re-targeting when it sees an ErrorCodes.NotMaster error response
* from a command even if "not master" doesn't appear in the message.
* @tags: [requires_replication]
*/
(function() {
"use strict";
// Set the refresh period to 10 min to rule out races
_setShellFailPoint({
configureFailPoint: "modifyReplicaSetMonitorDefaultRefreshPeriod",
mode: "alwaysOn",
data: {
period: 10 * 60,
},
});
const rst = new ReplSetTest({
nodes: 3,
nodeOptions: {
setParameter:
{"failpoint.respondWithNotPrimaryInCommandDispatch": tojson({mode: "alwaysOn"})}
}
});
rst.startSet();
rst.initiate();
const directConn = rst.getPrimary();
const rsConn = new Mongo(rst.getURL());
assert(rsConn.isReplicaSetConnection(),
"expected " + rsConn.host + " to be a replica set connection string");
const[secondary1, secondary2] = rst.getSecondaries();
function stepDownPrimary(rst) {
const awaitShell = startParallelShell(
() => assert.commandWorked(db.adminCommand({replSetStepDown: 60, force: true})),
directConn.port);
// We wait for the primary to transition to the SECONDARY state to ensure we're waiting
// until after the parallel shell has started the replSetStepDown command and the server is
// paused at the failpoint.Do not attempt to reconnect to the node, since the node will be
// holding the global X lock at the failpoint.
const reconnectNode = false;
rst.waitForState(directConn, ReplSetTest.State.SECONDARY, null, reconnectNode);
return awaitShell;
}
const failpoint = "stepdownHangBeforePerformingPostMemberStateUpdateActions";
assert.commandWorked(
directConn.adminCommand({configureFailPoint: failpoint, mode: "alwaysOn"}));
const awaitShell = stepDownPrimary(rst);
// Wait for a new primary to be elected and agreed upon by nodes.
rst.getPrimary();
rst.awaitNodesAgreeOnPrimary();
// DBClientRS will continue to send command requests to the node it believed to be primary even
// after it stepped down so long as it hasn't closed its connection.
assert.commandFailedWithCode(rsConn.getDB("test").runCommand({create: "mycoll"}),
ErrorCodes.NotMaster);
// However, once the server responds back with a ErrorCodes.NotMaster error, DBClientRS will
// cause the ReplicaSetMonitor to attempt to discover the current primary.
assert.commandWorked(rsConn.getDB("test").runCommand({create: "mycoll"}));
try {
assert.commandWorked(directConn.adminCommand({configureFailPoint: failpoint, mode: "off"}));
} catch (e) {
if (!isNetworkError(e)) {
throw e;
}
// We ignore network errors because it's possible that depending on how quickly the server
// closes connections that the connection would get closed before the server has a chance to
// respond to the configureFailPoint command with ok=1.
}
awaitShell();
rst.stopSet();
})();
|