WTCheckpointThread take up to 10 minutes on Mongo 4.2

Hello folks,

I have a 16TB Database in production (AWS) currently running on 3.6 and I want to upgrade it to 4.2. The issue I am having is related to the WiredTiger Checkpoints (WTCheckpointThread), those are taking very long time in 4.2 version. I noticed that because when running a daily snapshot backup on the Secondary DB it takes up to 10 minutes to complete the db.fsyncLock() so I checked the DB log and found those WiredTiger messages. Here is one example:
2020-07-28T03:37:28.690+0000 I STORAGE [WTCheckpointThread] WiredTiger message [1595907448:690341][35550:0x7f08d4e8b700], WT_SESSION.checkpoint: Checkpoint ran for 609 seconds and wrote: 418375 pages (26402 MB)

As I said before, I have the same DB size with the exact same DB resources and configurations and don’t have this issue there so I want to understand why this is happening or if there is something specific for 4.2 version that is causing this.

Can you please help.


I don

1 Like

I am having similar issues with multiple checkpoints running for multiple hours in 4.2