[SERVER-32049] significant performance degradation on mongo server with ~140K collections Created: 21/Nov/17  Updated: 06/Dec/22  Resolved: 12/Feb/18

Status: Closed
Project: Core Server
Component/s: None
Affects Version/s: 3.2.12
Fix Version/s: None

Type: Bug Priority: Critical - P2
Reporter: Avraham Apelbaum Assignee: Backlog - Triage Team
Resolution: Incomplete Votes: 0
Labels: collections, wiredtiger
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified

Assigned Teams:
Server Triage
Participants:

 Description   

Hi,
We have a mongo server with ~140K collections. After restart or new node sync all file descriptors of all collections stay open, which is normal, what is abnormal is that the app suffered from a severe performance degradation until those FDs where closed. Is there any limitation regarding number of collections.

Another question: By default mongo closes FDs after 28hrs by default, except for the startup options `file_manager=(close_idle_time=300)` is there any "dynamic" parameter . we can use ?

Or maximum number of open files is 265000

Br



 Comments   
Comment by Ramon Fernandez Marina [ 12/Feb/18 ]

tapuhi, we haven't heard back from you for a long time, so I'm going to close this ticket. If this is still an issue for you please feel free to comment here.

Thanks,
Ramón.

Comment by Bruce Lucas (Inactive) [ 21/Nov/17 ]

Hi Avraham,

We have made improvements in 3.4 in workloads with a large number of tables (see WT-3329). Would you be able to test your workload on 3.4.9 or later?

You can change the WiredTiger engine configuration dynamically using setParameter, e.g.

db.runCommand({setParameter: 1, wiredTigerEngineRuntimeConfig: "close_idle_time=300"})

Bruce

Comment by Avraham Apelbaum [ 21/Nov/17 ]

I forgot to mention we use WireTiger as storage engine.

Generated at Thu Feb 08 04:29:01 UTC 2024 using Jira 9.7.1#970001-sha1:2222b88b221c4928ef0de3161136cc90c8356a66.