[SERVER-16594] geoNear performance on low-end machine Created: 18/Dec/14  Updated: 09/Jan/15  Resolved: 09/Jan/15

Status: Closed
Project: Core Server
Component/s: Geo, Performance, Querying
Affects Version/s: None
Fix Version/s: None

Type: Bug Priority: Major - P3
Reporter: guipulsar Assignee: Siyuan Zhou
Resolution: Done Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified

Operating System: ALL
Participants:

 Description   

I also noticed a slow using Geonear with 2dsphere index,
but what is slow, im pretty new with mongo so imnot sure ,
here is the context:
i make a full scan over 35000 item collections with no limit result,
it takes about 5s on a 1core vm google cloud machine.
If its not a normal things (i guess) , i can provide you the collection for reproduce pb;
We just discover mongo and love it but this limitation is a serious blocker for us project.

db.runCommand({ geoNear: "tablegui", near: [2.298800, 48.854355], spherical: true, distanceMultiplier: 6371 , maxDistance:2000/6371 , limit:37000 });
 
"stats" :
{ "nscanned" : NumberLong(810310), "objectsLoaded" : NumberLong(810310), "avgDistance" : 325.45877488880984, "maxDistance" : 701.9530252234592, "time" : 5511 } , "ok" : 1



 Comments   
Comment by Siyuan Zhou [ 05/Jan/15 ]

Hey pulsar,

As far as I know, Google Cloud provides different kinds of 1 core VM with various capability, from shared CPU to a single hyperthread. Since database is memory (for cache) and I/O intensive and geo query needs more computation particularly, the performance sounds reasonable to me. If it's far from you expectation, the first step would be to understand the bottleneck in the system. Feel free to come back if you get more information that indicates a performance bug in MongoDB.

Thanks,
Siyuan

Generated at Thu Feb 08 03:41:36 UTC 2024 using Jira 9.7.1#970001-sha1:2222b88b221c4928ef0de3161136cc90c8356a66.