[SERVER-66131] Annotate commands sent with field indicating if the target rs is a CSRS Created: 02/May/22 Updated: 31/May/22 Resolved: 31/May/22 |
|
| Status: | Closed |
| Project: | Core Server |
| Component/s: | None |
| Affects Version/s: | None |
| Fix Version/s: | None |
| Type: | Task | Priority: | Major - P3 |
| Reporter: | Andrew Witten (Inactive) | Assignee: | Andrew Witten (Inactive) |
| Resolution: | Won't Do | Votes: | 0 |
| Labels: | sharding-nyc-subteam2-catalog-poc | ||
| Remaining Estimate: | Not Specified | ||
| Time Spent: | Not Specified | ||
| Original Estimate: | Not Specified | ||
| Participants: | |
| Story Points: | 3 |
| Description |
|
For PM-2290, there is now a designated data shard (the catalog shard) that also serves as the CSRS for the cluster. An issue that we are running into is that commands that enter the catalog shard should be handled differently depending on whether or not the command was sent to the catalog shard because it is the CSRS or because it is a data shard. Commands sent to the catalog shard are sent for one of two reasons. 1) Because it's the CSRS or 2) because it's a shard server. For the purpose of debugging we should annotate commands with this reason. |
| Comments |
| Comment by Andrew Witten (Inactive) [ 31/May/22 ] |
|
I no longer think this is useful so I am closing it as won't do. |
| Comment by Andrew Witten (Inactive) [ 03/May/22 ] |
|
I'm not sure if it needs to be this way, but aren't lots of things processed differently? We seem to register a bunch of different OpObservers and PrimaryOnlyServices on the CSRS vs on a shard server. I'm not really envisioning this as a change that would make it into final code, but more just for debugging the spike. |
| Comment by Kaloian Manassiev [ 03/May/22 ] |
|
What is the difference in handling that needs to be accounted for other than perhaps that such commands read with read concern majority and require after or atClusterTime? |