[SERVER-34436] Workloads to apply transaction snapshot cache pressure Created: 12/Apr/18  Updated: 29/Oct/23  Resolved: 02/Jul/18

Status: Closed
Project: Core Server
Component/s: Storage
Affects Version/s: None
Fix Version/s: 4.1.1

Type: Task Priority: Major - P3
Reporter: Dianna Hohensee (Inactive) Assignee: Dianna Hohensee (Inactive)
Resolution: Fixed Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified

Issue Links:
Depends
depends on SERVER-31767 Provide a window of snapshot history ... Closed
Related
is related to WT-4155 Snapshot history window bug: "evictio... Closed
is related to SERVER-35601 on startup setting of server paramete... Closed
is related to SERVER-35602 Add support to benchRun for atCluster... Closed
is related to SERVER-35603 add failpoint to prevent dynamic snap... Closed
Backwards Compatibility: Fully Compatible
Sprint: Storage NYC 2018-06-18, Storage NYC 2018-07-02, Storage NYC 2018-07-16
Participants:

 Description   

From the design:

"Test a varying workload that first applies cache pressure (many write transactions with one outstanding incomplete read transaction), then floods a server with PIT requests, then applies cache pressure again."

SERVER-31767 is introducing a snapshot history window within which a mongod guarantees service of reads, along with a few new server parameter knobs to handle cache pressure. Cache pressure can build as we keep too many snapshots open and the cache and server become immobilized, so we must test out configurations.

PIT requests are currently using the stable_timestamp, so this ticket may depend on further progress of the global PIT reads project, which depends on SERVER-31767.



 Comments   
Comment by Dianna Hohensee (Inactive) [ 14/Jun/18 ]

Turning this into a master ticket with linked tasks.

Generated at Thu Feb 08 04:36:41 UTC 2024 using Jira 9.7.1#970001-sha1:2222b88b221c4928ef0de3161136cc90c8356a66.