-
Type:
Question
-
Resolution: Duplicate
-
Priority:
Minor - P4
-
None
-
Affects Version/s: 2.4.14
-
Component/s: Aggregation Framework
-
None
Sometimes, sh.status() returns incomplete result.
mongos> sh.status()
Tue Sep 1 21:34:25 uncaught exception: group command failed: {
"retval" : [
{
"shard" : "s0",
"nChunks" : 7075
},
{
"shard" : "s1",
"nChunks" : 1806
},
{
"shard" : "s2",
"nChunks" : 600
},
{
"shard" : "s3",
"nChunks" : 8024
},
{
"shard" : "s4",
"nChunks" : 8024
}
],
"count" : 25529,
"keys" : 5,
"errmsg" : "exception: JavaScript execution terminated",
"code" : 16721,
"ok" : 0
}
At this time, config server print this message to it's log file.
Tue Sep 1 20:20:32.837 [conn11373] command config.$cmd command: { group: { cond: { ns: "db1.coll1" }, key: { shard: 1.0 }, initial: { nChunks: 0.0 }, ns: "chunks", $reduce: function ( doc , out ){ out.nChunks++; } } } ntoreturn:1 keyUpdates:0 numYields: 4 locks(micros) r:2255642 reslen:305 1415ms
Tue Sep 1 20:20:33.082 [conn11373] JavaScript execution terminated
And according to the source code (src/mongo/db/commands/group.cpp), Looks like aggregation function has 100ms timeout. (I am not sure this is exact cause of incomplete result of sh.status())
s->setObject( "$initial" , initial , true );
s->exec( "$reduce = " + reduceCode , "$group reduce setup" , false , true , true , 100 );
s->exec( "$arr = [];" , "$group reduce setup 2" , false , true , true , 100 );
ScriptingFunction f = s->createFunction(
"function(){ "
" if ( $arr[n] == null ){ "
" next = {}; "
" Object.extend( next , $key ); "
" Object.extend( next , $initial , true ); "
" $arr[n] = next; "
" next = null; "
" } "
" $reduce( obj , $arr[n] ); "
"}" );
Is this something to do with SERVER-17012.
And Can I change sh.status command(aggreation framework) timeout.
We have 3 collections and two of them have about 25000 chunks each and the other one has a few chunks.
- duplicates
-
SERVER-13985 printShardingStatus uses group/JS
-
- Closed
-