[SERVER-36969] initial_sync_wt_cache_full.js takes too long to complete on slow hosts Created: 31/Aug/18  Updated: 29/Oct/23  Resolved: 05/Sep/18

Status: Closed
Project: Core Server
Component/s: Storage
Affects Version/s: 4.1 Desired
Fix Version/s: 3.6.9, 4.0.3, 4.1.3

Type: Bug Priority: Major - P3
Reporter: Benety Goh Assignee: Benety Goh
Resolution: Fixed Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified

Issue Links:
Backports
Depends
Backwards Compatibility: Fully Compatible
Operating System: ALL
Backport Requested:
v4.0, v3.6
Sprint: Storage NYC 2018-09-10
Participants:
Linked BF Score: 17

 Description   

On some slow machines, the oplog application phase of initial sync in initial_sync_wt_cache_full.js could take longer than the default timeout in ReplSetTest to complete. We can use a longer timeout to make this test more robust on slow hosts.



 Comments   
Comment by Githook User [ 18/Sep/18 ]

Author:

{'name': 'Benety Goh', 'email': 'benety@mongodb.com', 'username': 'benety'}

Message: SERVER-36969 reduce number of updates in wt cache tests from 1000 to 500

(cherry picked from commit 5e6c3d0b10a0197a4df6df408d3110d87b67638a)
Branch: v3.6
https://github.com/mongodb/mongo/commit/9f0fd6a76d91f4b55cb07bcab9f7f23407877e8c

Comment by Githook User [ 06/Sep/18 ]

Author:

{'name': 'Benety Goh', 'email': 'benety@mongodb.com', 'username': 'benety'}

Message: SERVER-36969 reduce number of updates in wt cache tests from 1000 to 500

(cherry picked from commit 5e6c3d0b10a0197a4df6df408d3110d87b67638a)
Branch: v4.0
https://github.com/mongodb/mongo/commit/4f3d84db7029a14badb675f26238871e9ec70142

Comment by Githook User [ 05/Sep/18 ]

Author:

{'name': 'Benety Goh', 'email': 'benety@mongodb.com', 'username': 'benety'}

Message: SERVER-36969 reduce number of updates in wt cache tests from 1000 to 500
Branch: master
https://github.com/mongodb/mongo/commit/5e6c3d0b10a0197a4df6df408d3110d87b67638a

Comment by Benety Goh [ 01/Sep/18 ]

max.hirschhorn, it's probably just I/O on the slow machines for the variant where we observed this test failure. The other cache pressure tests have comparable completion times. I'm looking into reducing the number of updates, currently 1000 for each of the two 10 MB documents, to a more manageable number.

Comment by Max Hirschhorn [ 01/Sep/18 ]

benety.goh, do we know why the initial_sync_wt_cache_full.js test takes so long to finish the oplog application phase? What resource is the bottleneck?

Generated at Thu Feb 08 04:44:36 UTC 2024 using Jira 9.7.1#970001-sha1:2222b88b221c4928ef0de3161136cc90c8356a66.