You are viewing a plain text version of this content. The canonical link for it is here.
Posted to notifications@accumulo.apache.org by "DomGarguilo (via GitHub)" <gi...@apache.org> on 2026/03/12 17:09:18 UTC
[PR] WIP - Refactor External Compactions monitor view to use snapshot-backed Top 50 data [accumulo]
DomGarguilo opened a new pull request, #6211:
URL: https://github.com/apache/accumulo/pull/6211
Fixes #6174
This PR:
* moves EC monitor data gathering into `InformationFetcher`/`SystemInformation` snapshot flow
* serves the active compaction data from top 50 longest running compactions view
* removes the old/competing `rest-v2/compactions/detail` path
* makes no intended changes to the active compactions page and just refactors the data collection on the backend
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: notifications-unsubscribe@accumulo.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
Re: [PR] WIP - Refactor External Compactions monitor view to use snapshot-backed Top 50 data [accumulo]
Posted by "DomGarguilo (via GitHub)" <gi...@apache.org>.
DomGarguilo merged PR #6211:
URL: https://github.com/apache/accumulo/pull/6211
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: notifications-unsubscribe@accumulo.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
Re: [PR] WIP - Refactor External Compactions monitor view to use snapshot-backed Top 50 data [accumulo]
Posted by "DomGarguilo (via GitHub)" <gi...@apache.org>.
DomGarguilo commented on PR #6211:
URL: https://github.com/apache/accumulo/pull/6211#issuecomment-4048562240
> I don't think we should keep the full list of external compactions in memory in the Monitor for two reasons
> 1. It could eat up a lot of memory
> 2. The information may not be useful or accessed for very quick compactions
>
> Instead I think we can do the following to show the longest running compactions in the cluster:
>
> 1. Move the `TimeOrderedRunningCompactionSet` object from the Coordinator to the Monitor
> 2. As the Monitor is building the next version of the SystemInformation object, populate the `TimeOrderedRunningCompactionSet` set with information from each Compactors `getActiveCompactions` result.
> 3. Allow the size of `TimeOrderedRunningCompactionSet` to be configured by the user so they can decide how much memory to use for this feature.
> 4. Provide views of the data in `TimeOrderedRunningCompactionSet` by table, by resource group, and a global view.
_Originally posted by @dlmarion in [#6174](https://github.com/apache/accumulo/issues/6174#issuecomment-4041847924)_
This branch still keeps the full running compactions set internally so ec/details can continue to work. A follow-on could drop that details support and keep only a Top N set in memory.
Another follow-on would be to move Top N collection fully into the Monitor by polling each Compactor while building SystemInformation, instead of asking the manager/coordinator for that info. That would reduce reliance on the manager/coordinator, but would add more collection logic complexity so we would need to consider that.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: notifications-unsubscribe@accumulo.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org