[SERVER-10327] too much logging when secondaries slow in chunk migrations src/mongo/s/d_migrate.cpp Created: 25/Jul/13 Updated: 06/Dec/22 Resolved: 23/Aug/18 |
|
| Status: | Closed |
| Project: | Core Server |
| Component/s: | Sharding |
| Affects Version/s: | 2.2.3 |
| Fix Version/s: | None |
| Type: | Bug | Priority: | Minor - P4 |
| Reporter: | AHL Linux Support | Assignee: | [DO NOT USE] Backlog - Sharding Team |
| Resolution: | Done | Votes: | 0 |
| Labels: | ChunkMigrationRefactor | ||
| Remaining Estimate: | Not Specified | ||
| Time Spent: | Not Specified | ||
| Original Estimate: | Not Specified | ||
| Issue Links: |
|
||||||||
| Assigned Teams: |
Sharding
|
||||||||
| Operating System: | ALL | ||||||||
| Steps To Reproduce: | Rebuild many secondaries in a replica set and stress the primary. |
||||||||
| Participants: | |||||||||
| Description |
|
Between lines 1565 and 1580, the message "secondaries having hard time keeping up with migrate" can be output to the log every 20ms for the duration of i==100 to i==maxIterations. This causes log spam. In addition,the message is neither intuitive nor helpful. |
| Comments |
| Comment by Kaloian Manassiev [ 23/Aug/18 ] |
|
With the changes to the migration machinery in 3.4, this message is no longer printed. |
| Comment by Dave Muysson [ 09/Jun/16 ] |
|
I know this is an old ticket, but our logging ingest solution almost tipped over because of the shear volume of this message coming out of MongoD. We are running 3.0.12 and initiated a resync of our Secondaries, afterwhich one of the shards started logging this every 20ms. If this is supposed to be informational, could it at least be placed under "I" instead of "W" in the log output? |
| Comment by Stennie Steneker (Inactive) [ 13/Aug/13 ] |
|
Hi, Thank you for the feedback. This warning message is intended to provide some information in the event the secondaries may be having challenges keeping up with migrations prior to the failed state of "secondary can't keep up with migrate". Currently this is emitted on each loop after there have been more than 100 iterations, so we could look at reducing the frequency to minimize the log spam. Regards, |