[
https://issues.apache.org/jira/browse/GOBBLIN-1787?focusedWorklogId=845699&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-845699
]
ASF GitHub Bot logged work on GOBBLIN-1787:
-------------------------------------------
Author: ASF GitHub Bot
Created on: 15/Feb/23 18:47
Start Date: 15/Feb/23 18:47
Worklog Time Spent: 10m
Work Description: homatthew opened a new pull request, #3645:
URL: https://github.com/apache/gobblin/pull/3645
Dear Gobblin maintainers,
Please accept this PR. I understand that it will not be reviewed until I
have checked off all the steps below!
Let's improve the interface for state stores that Gobblin uses to storing
watermarks and various other pieces of state. State stores like zookeeper have
native methods for batch deletes that we cannot take advantage of with the
current implementation.
This is useful for cleaning up no longer used watermarks in a state store.
The size of a state store has physical limitations depending on the underlying
implementation (e.g. ZK supports roughly ~10k tables per state store). And even
migrating to a new state store requires a lot of operational burden / doesn't
clean up the old state store.
Design:
- Add a default method to the interface for backward compatibility in the
OSS community
- Use this method in the watermark based storage
- Call batch delete in zk state store
Alternate designs:
- Create a new interface that extends the interface `StateStore` except it
has a delete multiple tables method. The default approach is a much cleaner
approach in terms of backwards compatibility / amount of code needed.
- Add batch deletes as part of the watermark based storage, but this would
be a hack since we'd just be calling delete with multiple threads
### JIRA
- [X] My PR addresses the following [Gobblin
JIRA](https://issues.apache.org/jira/browse/GOBBLIN/) issues and references
them in the PR title. For example, "[GOBBLIN-1787] My Gobblin PR"
- https://issues.apache.org/jira/browse/GOBBLIN-1787
### Description
- [X] Here are some details about my PR, including screenshots (if
applicable):
### Tests
- [X] My PR adds the following unit tests __OR__ does not need testing for
this extremely good reason:
Used batch delete of a singleton collection in UT
### Commits
- [ ] My commits all reference JIRA issues in their subject lines, and I
have squashed multiple commits if they address the same issue. In addition, my
commits follow the guidelines from "[How to write a good git commit
message](http://chris.beams.io/posts/git-commit/)":
1. Subject is separated from body by a blank line
2. Subject is limited to 50 characters
3. Subject does not end with a period
4. Subject uses the imperative mood ("add", not "adding")
5. Body wraps at 72 characters
6. Body explains "what" and "why", not "how"
Issue Time Tracking
-------------------
Worklog Id: (was: 845699)
Remaining Estimate: 0h
Time Spent: 10m
> Update watermark based state store to be able to delete watermarks
> ------------------------------------------------------------------
>
> Key: GOBBLIN-1787
> URL: https://issues.apache.org/jira/browse/GOBBLIN-1787
> Project: Apache Gobblin
> Issue Type: Improvement
> Reporter: Matthew Ho
> Priority: Major
> Time Spent: 10m
> Remaining Estimate: 0h
>
> The state store based watermark is unable to delete watermarks for cleanup /
> maintenance. This creates an accumulation in stores like ZK or mysql, which
> have a physical limit on the number of tables in a state store
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
