[SERVER-44187] High I/O load in Secondary node Created: 24/Oct/19  Updated: 27/Oct/23  Resolved: 11/Nov/19

Status: Closed
Project: Core Server
Component/s: WiredTiger
Affects Version/s: 3.2.10
Fix Version/s: None

Type: Bug Priority: Major - P3
Reporter: miao yimin Assignee: Dmitry Agranat
Resolution: Works as Designed Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified

Attachments: JPEG File primary iops.jpg     JPEG File primary mongostat.jpg     JPEG File secondary iops.jpg     JPEG File secondary mongostat.jpg    
Backwards Compatibility: Fully Compatible
Operating System: ALL
Steps To Reproduce:

(1)Keep the database operation scenario unchanged

(2)Watch the primary iops

(3)Watch the secondary iops
Observing the main iops situation

Participants:

 Description   

In the scenario of continuous insertion and update, the role of mongodb is changed on the same host, and iostat is used to monitor the performance of io. It is found that the value of %util is much lower than that of seconary while in role primary.



 Comments   
Comment by Dmitry Agranat [ 24/Oct/19 ]

Hi 879811746@qq.com,

In general, uneven i/o during the write workload between the Primary and Secondaries can be attributed to different reasons, for example, due to the way we are flushing on every replication batch to ensure durability of each batch.

We'd be happy to have a closer look once your cluster is running on one of our latest MongoDB versions (as of today MongoDB 4.2.1 ) as MongoDB 3.2 has reached EOL on September 2018.

Thanks,
Dima

Generated at Thu Feb 08 05:05:16 UTC 2024 using Jira 9.7.1#970001-sha1:2222b88b221c4928ef0de3161136cc90c8356a66.