[JAVA-3756] When continued polling via the java driver using a resume token not in oplog, leads to mongos's crashing Created: 05/Jun/20  Updated: 23/Jun/20  Resolved: 23/Jun/20

Status: Closed
Project: Java Driver
Component/s: None
Affects Version/s: 3.6.4
Fix Version/s: None

Type: Bug Priority: Major - P3
Reporter: Anand Menon Assignee: Rachelle Palmer
Resolution: Done Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified


 Description   

Leads to:

  1. build up of cursor counts in mongod
  2. slow get-more queries
  3. high CPU usage in mongod
  4. high memory usage in mongos

Observations:

  1. The maxAwaitTimeMS set in the watch command not honoured by queries in mongod - Does not return result nor empty within timeout 
  2. The cursor.getnext operation creating new connections and cursors behind the scenes
  3. If resume token is not matching any data, what is filling up in mongos memory?


 Comments   
Comment by Rachelle Palmer [ 06/Jun/20 ]

Hi there, thank you for reaching out. As this sounds like a support issue, we'd like to point to resources to get this questioned answered:

  • our MongoDB support portal, located at support.mongodb.com
  • our MongoDB community portal, located here
  • for Atlas users, there is a built in chat feature in the UI, accessible in the lower right hand corner (it looks like a chat icon). Chat is staffed 24/7 by support personnel.

Folks at any of those can assist in diagnosing and/or reproducing the issue.

Rachelle

Generated at Thu Feb 08 09:00:21 UTC 2024 using Jira 9.7.1#970001-sha1:2222b88b221c4928ef0de3161136cc90c8356a66.