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
85
86
87
88
89
90
|
/**
* Verify the FTDC metrics for reactor threads.
*
* @tags: [
* requires_sharding,
* ]
*/
load('jstests/libs/fail_point_util.js');
load('jstests/libs/ftdc.js');
load('jstests/libs/parallelTester.js');
(function() {
'use strict';
function getDiagnosticData(mongos) {
let db = mongos.getDB('admin');
const stats = verifyGetDiagnosticData(db);
assert(stats.hasOwnProperty('networkInterfaceStats'));
return stats.networkInterfaceStats;
}
const numThreads = 50;
const ftdcPath = MongoRunner.toRealPath('ftdc');
const st = new ShardingTest({
shards: 1,
mongos: {
s0: {setParameter: {diagnosticDataCollectionDirectoryPath: ftdcPath}},
}
});
// Block operations after they acquire a connection.
const fp = 'networkInterfaceHangCommandsAfterAcquireConn';
let res = assert.commandWorked(st.s.adminCommand({configureFailPoint: fp, mode: 'alwaysOn'}));
// Run some operations and have them all blocked after acquiring a connection.
let threads = [];
for (var t = 0; t < numThreads; t++) {
let thread = new Thread(function(connStr, t) {
var conn = new Mongo(connStr);
conn.getDB('test').runCommand({insert: 'test', documents: [{counter: t}]});
}, st.s.host, t);
threads.push(thread);
thread.start();
}
// Wait for the reactor thread to block on the fail point.
// There is only one reactor thread for each network interface, and we can block that thread on `fp`
// by establishing new egress connections. The threads defined earlier trigger establishment of
// egress connections. Once the reactor thread is blocked, the number of times the server enters
// `fp` will no longer advance.
for (var t = 1; t <= numThreads; t++) {
assert.neq(t, numThreads, "Failed to make the reactor thread block on the fail point");
try {
jsTestLog(`Checking fail point for times entered: ${res.count + t} ...`);
assert.commandWorked(st.s.adminCommand({
waitForFailPoint: fp,
timesEntered: res.count + t,
maxTimeMS: 10000, // A large timeout to mitigate scheduling issues on slow machines.
}));
} catch (ex) {
assert.commandFailedWithCode(ex, ErrorCodes.MaxTimeMSExpired);
jsTestLog('The reactor thread should be blocked now!');
break;
}
}
// Introduce some delay before disabling the fail point and unblocking the operation.
sleep(5000);
assert.commandWorked(st.s.adminCommand({configureFailPoint: fp, mode: 'off'}));
for (var t = 0; t < numThreads; t++) {
threads[t].join();
}
jsTestLog("Verifying FTDC metrics ...");
assert.soon(() => {
const metrics = getDiagnosticData(st.s);
// We expect at least one of the tasks scheduled on reactor threads to have a long run time.
let longRunningTasks = 0; // Tasks with a run time > 1 sec.
for (const instance in metrics) {
if (!metrics[instance].hasOwnProperty('runTime'))
continue; // Filter out FTDC metadata.
longRunningTasks += metrics[instance]['runTime']['1000ms+'];
}
return longRunningTasks >= 1;
}, 'Expected to find at least one long running task', 30000);
st.stop();
})();
|