You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@solr.apache.org by "Houston Putman (Jira)" <ji...@apache.org> on 2021/10/29 17:20:00 UTC

[jira] [Resolved] (SOLR-15681) Customization of S3 client retry/throttling logic

     [ https://issues.apache.org/jira/browse/SOLR-15681?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Houston Putman resolved SOLR-15681.
-----------------------------------
    Fix Version/s: 8.11
         Assignee: Houston Putman
       Resolution: Workaround

SOLR-15707 adds 2 options for S3Repositories:

* Set the profile for the S3Repository, which lets users set the RetryMode and maxAttempts. These options can also be set globally via environment variables.
* {{s3.retries.disable}} to disable all retries for that repository.

With these options available, I think this ticket can be considered "done" for now. There are other retry customization options available in the S3 client, but this seems like a good MVP.

> Customization of S3 client retry/throttling logic
> -------------------------------------------------
>
>                 Key: SOLR-15681
>                 URL: https://issues.apache.org/jira/browse/SOLR-15681
>             Project: Solr
>          Issue Type: Improvement
>          Components: contrib - S3 Repository
>            Reporter: Houston Putman
>            Assignee: Houston Putman
>            Priority: Major
>             Fix For: 8.11
>
>
> Currently there are very few configuration options for users to customize how the s3-repository module interacts with S3.
> One such option that would be very beneficial, especially given how many files Solr backups can use, would be retry and throttling logic. The AWS client provides [a few options|https://docs.aws.amazon.com/sdk-for-java/latest/developer-guide/using.html#using-retries] to customize the number of retries, and backoff logic, when requests do not succeed.
> We don't want to give users a 1000 options to configure the S3 client in the solr.xml, but we can definitely give a few popular options that would help optimize for their use cases. Retries and throttling backoff logic seem like 2 good options to start with.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@solr.apache.org
For additional commands, e-mail: issues-help@solr.apache.org