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 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304
|
// -------------------------
// SPLITVECTOR TEST UTILS
// -------------------------
// -------------------------
// assertChunkSizes verifies that a given 'splitVec' divides the 'test.jstest_splitvector'
// collection in 'maxChunkSize' approximately-sized chunks. Its asserts fail otherwise.
// @param splitVec: an array with keys for field 'x'
// e.g. [ { x : 1927 }, { x : 3855 }, ...
// @param numDocs: domain of 'x' field
// e.g. 20000
// @param maxChunkSize is in MBs.
//
assertChunkSizes = function ( splitVec , numDocs , maxChunkSize , msg ){
splitVec = [{ x: -1 }].concat( splitVec );
splitVec.push( { x: numDocs+1 } );
for ( i=0; i<splitVec.length-1; i++) {
min = splitVec[i];
max = splitVec[i+1];
size = db.runCommand( { datasize: "test.jstests_splitvector" , min: min , max: max } ).size;
// It is okay for the last chunk to be smaller. A collection's size does not
// need to be exactly a multiple of maxChunkSize.
if ( i < splitVec.length - 2 )
assert.close( maxChunkSize , size , "A"+i , -3 );
else
assert.gt( maxChunkSize , size , "A"+i , msg + "b" );
}
}
// Takes two documents and asserts that both contain exactly the same set of field names.
// This is useful for checking that splitPoints have the same format as the original key pattern,
// even when sharding on a prefix key.
// Not very efficient, so only call when # of field names is small
var assertFieldNamesMatch = function( splitPoint , keyPattern ){
for ( var p in splitPoint ) {
if( splitPoint.hasOwnProperty( p ) ) {
assert( keyPattern.hasOwnProperty( p ) , "property " + p + " not in keyPattern" );
}
}
for ( var p in keyPattern ) {
if( keyPattern.hasOwnProperty( p ) ){
assert( splitPoint.hasOwnProperty( p ) , "property " + p + " not in splitPoint" );
}
}
}
// -------------------------
// TESTS START HERE
// -------------------------
f = db.jstests_splitvector;
f.drop();
// -------------------------
// Case 1: missing parameters
assert.eq( false, db.runCommand( { splitVector: "test.jstests_splitvector" } ).ok , "1a" );
assert.eq( false, db.runCommand( { splitVector: "test.jstests_splitvector" , maxChunkSize: 1} ).ok , "1b" );
// -------------------------
// Case 2: missing index
assert.eq( false, db.runCommand( { splitVector: "test.jstests_splitvector" , keyPattern: {x:1} , maxChunkSize: 1 } ).ok , "2");
// -------------------------
// Case 3: empty collection
f.ensureIndex( { x: 1} );
assert.eq( [], db.runCommand( { splitVector: "test.jstests_splitvector" , keyPattern: {x:1} , maxChunkSize: 1 } ).splitKeys , "3");
// -------------------------
// Case 4: uniform collection
f.drop();
f.ensureIndex( { x: 1 } );
var case4 = function() {
// Get baseline document size
filler = "";
while( filler.length < 500 ) filler += "a";
f.save( { x: 0, y: filler } );
docSize = db.runCommand( { datasize: "test.jstests_splitvector" } ).size;
assert.gt( docSize, 500 , "4a" );
// Fill collection and get split vector for 1MB maxChunkSize
numDocs = 20000;
for( i=1; i<numDocs; i++ ){
f.save( { x: i, y: filler } );
}
db.getLastError();
res = db.runCommand( { splitVector: "test.jstests_splitvector" , keyPattern: {x:1} , maxChunkSize: 1 } );
// splitVector aims at getting half-full chunks after split
factor = 0.5;
assert.eq( true , res.ok , "4b" );
assert.close( numDocs*docSize / ((1<<20) * factor), res.splitKeys.length , "num split keys" , -1 );
assertChunkSizes( res.splitKeys , numDocs, (1<<20) * factor , "4d" );
for( i=0; i < res.splitKeys.length; i++ ){
assertFieldNamesMatch( res.splitKeys[i] , {x : 1} );
}
}
case4();
// -------------------------
// Case 5: limit number of split points
f.drop();
f.ensureIndex( { x: 1 } );
var case5 = function() {
// Fill collection and get split vector for 1MB maxChunkSize
numDocs = 10000;
for( i=1; i<numDocs; i++ ){
f.save( { x: i, y: filler } );
}
db.getLastError();
res = db.runCommand( { splitVector: "test.jstests_splitvector" , keyPattern: {x:1} , maxChunkSize: 1 , maxSplitPoints: 1} );
assert.eq( true , res.ok , "5a" );
assert.eq( 1 , res.splitKeys.length , "5b" );
for( i=0; i < res.splitKeys.length; i++ ){
assertFieldNamesMatch( res.splitKeys[i] , {x : 1} );
}
}
case5();
// -------------------------
// Case 6: limit number of objects in a chunk
f.drop();
f.ensureIndex( { x: 1 } );
var case6 = function() {
// Fill collection and get split vector for 1MB maxChunkSize
numDocs = 10000;
for( i=1; i<numDocs; i++ ){
f.save( { x: i, y: filler } );
}
db.getLastError();
res = db.runCommand( { splitVector: "test.jstests_splitvector" , keyPattern: {x:1} , maxChunkSize: 1 , maxChunkObjects: 500} );
assert.eq( true , res.ok , "6a" );
assert.eq( 19 , res.splitKeys.length , "6b" );
for( i=0; i < res.splitKeys.length; i++ ){
assertFieldNamesMatch( res.splitKeys[i] , {x : 1} );
}
}
case6();
// -------------------------
// Case 7: enough occurances of min key documents to pass the chunk limit
// [1111111111111111,2,3)
f.drop();
f.ensureIndex( { x: 1 } );
var case7 = function() {
// Fill collection and get split vector for 1MB maxChunkSize
numDocs = 2100;
for( i=1; i<numDocs; i++ ){
f.save( { x: 1, y: filler } );
}
for( i=1; i<10; i++ ){
f.save( { x: 2, y: filler } );
}
db.getLastError();
res = db.runCommand( { splitVector: "test.jstests_splitvector" , keyPattern: {x:1} , maxChunkSize: 1 } );
assert.eq( true , res.ok , "7a" );
assert.eq( 2 , res.splitKeys[0].x, "7b");
for( i=0; i < res.splitKeys.length; i++ ){
assertFieldNamesMatch( res.splitKeys[i] , {x : 1} );
}
}
case7();
// -------------------------
// Case 8: few occurrances of min key, and enough of some other that we cannot split it
// [1, 22222222222222, 3)
f.drop();
f.ensureIndex( { x: 1 } );
var case8 = function() {
for( i=1; i<10; i++ ){
f.save( { x: 1, y: filler } );
}
numDocs = 2100;
for( i=1; i<numDocs; i++ ){
f.save( { x: 2, y: filler } );
}
for( i=1; i<10; i++ ){
f.save( { x: 3, y: filler } );
}
db.getLastError();
res = db.runCommand( { splitVector: "test.jstests_splitvector" , keyPattern: {x:1} , maxChunkSize: 1 } );
assert.eq( true , res.ok , "8a" );
assert.eq( 2 , res.splitKeys.length , "8b" );
assert.eq( 2 , res.splitKeys[0].x , "8c" );
assert.eq( 3 , res.splitKeys[1].x , "8d" );
for( i=0; i < res.splitKeys.length; i++ ){
assertFieldNamesMatch( res.splitKeys[i] , {x : 1} );
}
}
case8();
// -------------------------
// Case 9: splitVector "force" mode, where we split (possible small) chunks in the middle
//
f.drop();
f.ensureIndex( { x: 1 } );
var case9 = function() {
f.save( { x: 1 } );
f.save( { x: 2 } );
f.save( { x: 3 } );
db.getLastError();
assert.eq( 3 , f.count() );
print( f.getFullName() )
res = db.runCommand( { splitVector: f.getFullName() , keyPattern: {x:1} , force : true } );
assert.eq( true , res.ok , "9a" );
assert.eq( 1 , res.splitKeys.length , "9b" );
assert.eq( 2 , res.splitKeys[0].x , "9c" );
if ( db.runCommand( "isMaster" ).msg != "isdbgrid" ) {
res = db.adminCommand( { splitVector: "test.jstests_splitvector" , keyPattern: {x:1} , force : true } );
assert.eq( true , res.ok , "9a: " + tojson(res) );
assert.eq( 1 , res.splitKeys.length , "9b: " + tojson(res) );
assert.eq( 2 , res.splitKeys[0].x , "9c: " + tojson(res) );
for( i=0; i < res.splitKeys.length; i++ ){
assertFieldNamesMatch( res.splitKeys[i] , {x : 1} );
}
}
}
case9();
// -------------------------
// Repeat all cases using prefix shard key.
//
f.drop();
f.ensureIndex( { x: 1, y: 1 } );
case4();
f.drop();
f.ensureIndex( { x: 1, y: -1 , z : 1 } );
case4();
f.drop();
f.ensureIndex( { x: 1, y: 1 } );
case5();
f.drop();
f.ensureIndex( { x: 1, y: -1 , z : 1 } );
case5();
f.drop();
f.ensureIndex( { x: 1, y: 1 } );
case6();
f.drop();
f.ensureIndex( { x: 1, y: -1 , z : 1 } );
case6();
f.drop();
f.ensureIndex( { x: 1, y: 1 } );
case7();
f.drop();
f.ensureIndex( { x: 1, y: -1 , z : 1 } );
case7();
f.drop();
f.ensureIndex( { x: 1, y: 1 } );
case8();
f.drop();
f.ensureIndex( { x: 1, y: -1 , z : 1 } );
case8();
f.drop();
f.ensureIndex( { x: 1, y: 1 } );
case9();
f.drop();
f.ensureIndex( { x: 1, y: -1 , z : 1 } );
case9();
print("PASSED");
|