[ 
https://issues.apache.org/jira/browse/SOLR-8950?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15227892#comment-15227892
 ] 

Jan Høydahl commented on SOLR-8950:
-----------------------------------

+1 to "atomic rename & alias"

A bit outside the scope of this JIRA, but perhaps later we could then also add 
a convenience restore collection command which replaces an existing collection 
atomically, e.g. ({{&name=foo&replace=true}}). Under the hoods, it would do:
# Restore backup to {{foo.restored.20151224121300}} (time stamp of the backup)
# If foo exists and is a collection: atomic rename foo as 
{{foo.20160406092300}} (current timestamp), and add alias {{foo --> 
foo.20160406092300}}
# Atomic replace alias foo to point to {{foo.restored.20151224121300}}
# Optionally, rename old real collection as {{trash.foo.20160406092300}} 
(simulate a trashcan), so it is easy to delete it later, manually or through a 
"purge" command



> Provide an option to "disable" a specified Solr collection
> ----------------------------------------------------------
>
>                 Key: SOLR-8950
>                 URL: https://issues.apache.org/jira/browse/SOLR-8950
>             Project: Solr
>          Issue Type: New Feature
>          Components: SolrCloud
>    Affects Versions: 4.10.3, trunk
>            Reporter: Hrishikesh Gadre
>
> Currently Solr does not provide facility to "disable" read/write requests for 
> a specified collection. This is essential for certain admin operations (e.g. 
> restore collection data from an earlier snapshot). Without this support, we 
> need to bring down the Solr cluster during the snapshot recovery. In case of 
> multi-tenant Solr cluster, it is not always feasible to bring down the Solr 
> service since it can affect the availability of other collections too.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to