Databases
StackOverflow2013
Postsdbo
POmongodb update operation on shards very slow
 

Note that there are some explanatory texts on larger screens.

plurals
  1. POmongodb update operation on shards very slow
    primarykey
    data
    text
    <p>I have a little cluster which consists of 3 shards, and every shard is a replica set of 2 real nodes and 1 ARBITER. sharding is enabled on a collection, let's say <code>coll_words</code>, key is <code>title</code>.</p> <p>other info:</p> <ul> <li><p><code>rs.slaveOk()</code> is on and <code>balancer</code> is off. </p></li> <li><p>the primary shard of this collection contains around 13M documents, the other 2 secondary ones each contains around 5M.</p></li> <li><p>average ping between shards is below 1 millis. Pings from mongos to shard is about <code>3 Millis</code> with <code>'connPoolStats'</code> command.</p></li> </ul> <p>The typical <code>find()</code> response through a mongos like this (<code>update()</code> is slow too):</p> <pre><code>mongos&gt; db.coll_words.find({title:'りのフランス'}).explain() { "clusteredType" : "ParallelSort", "shards" : { "rsName/IP1:27017,IP2:27017" : [ { "cursor" : "BtreeCursor title_1", "isMultiKey" : false, "n" : 1, "nscannedObjects" : 1, "nscanned" : 1, "nscannedObjectsAllPlans" : 1, "nscannedAllPlans" : 1, "scanAndOrder" : false, "indexOnly" : false, "nYields" : 0, "nChunkSkips" : 0, "millis" : 0, "indexBounds" : { "title" : [ [ "りのフランス", "りのフランス" ] ] }, "server" : "SERVER:27017" } ] }, "cursor" : "BtreeCursor title_1", "n" : 1, "nChunkSkips" : 0, "nYields" : 0, "nscanned" : 1, "nscannedAllPlans" : 1, "nscannedObjects" : 1, "nscannedObjectsAllPlans" : 1, "millisShardTotal" : 0, "millisShardAvg" : 0, "numQueries" : 1, "numShards" : 1, "indexBounds" : { "title" : [ [ "りのフランス", "りのフランス" ] ] }, "millis" : 51 } </code></pre> <p>></p> <p>It's really confusing to me, why this <em>finally</em> costs <strong><code>"millis" : 51</code></strong>? </p> <p>Am I missing something? Any pointers would be great.</p> <p><strong>Update:</strong></p> <p>Pings from mongos to shard is about <code>3 Millis</code> with <code>'connPoolStats'</code> command.</p> <pre><code>"replicaSets" : { "rs0" : { "hosts" : [ { "addr" : "IP1:27017", "ok" : true, "ismaster" : false, "hidden" : false, "secondary" : true, "pingTimeMillis" : 3 }, { "addr" : "IP2:27017", "ok" : true, "ismaster" : true, "hidden" : false, "secondary" : false, "pingTimeMillis" : 2 } ], "master" : 1, "nextSlave" : 0 }, </code></pre>
    singulars
    1. This table or related slice is empty.
    1. This table or related slice is empty.
    1. PTQuestion
    1. USmaow
    1. USmaow
    plurals
    1. This table or related slice is empty.
    1. This table or related slice is empty.
    1. This table or related slice is empty.
    1. This table or related slice is empty.
    1. VO
      singulars
      1. POmongodb update operation on shards very slow
      1. This table or related slice is empty.
      1. VTUpMod
    1. COInvestigate the issue by creating a test, sharded, balanced, small collection and query it with different keys.
      singulars
      1. POmongodb update operation on shards very slow
      1. USDyin
    2. COYou say how long the pings between shards are - but that's irrelevant - how long is the ping from mongos to shard rsName? Clearly the query itself is nearly instant, but it's not getting to/from mongos very fast.
      singulars
      1. POmongodb update operation on shards very slow
      1. USAsya Kamsky
    3. COThanks for your reply. Pings from mongos to shards are nearly instant too, what did you mean exactly by `getting to/from mongos is not very fast`?
      singulars
      1. POmongodb update operation on shards very slow
      1. USmaow
 

Querying!

 
Guidance
A row detail

Detail views are divided into sections. All the information in the data section comes from columns in the selected row. The other sections display data from other, related rows.

Related data can be related in a to-one or a to-many fashion. Captions of data related in a to-many fashion link to a list view showing a filtered view of the table.

Try moving around until you find a non-empty to-many entry and click on the label to get to one. You can move back to the root by clicking on the database name in the header.

SQuiL has stopped working due to an internal error.

If you are curious you may find further information in the browser console, which is accessible through the devtools (F12).

Reload