cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jonathan Ellis (JIRA)" <j...@apache.org>
Subject [jira] Commented: (CASSANDRA-1075) Thrift interface call to query whether all live nodes agree on the schema id
Date Wed, 19 May 2010 06:14:53 GMT

    [ https://issues.apache.org/jira/browse/CASSANDRA-1075?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12869011#action_12869011
] 

Jonathan Ellis commented on CASSANDRA-1075:
-------------------------------------------

on 01,

minor points:

Math.max(DatabaseDescriptor.getRpcTimeout()-1000, 500)

why not just use DD.getRpcTimeout?  feels like trying to be too clever.

+        catch (InterruptedException ex) 
+        {
+        }

let's throw assertionerror for these.

Finally, the .info logs feel like they should be .debug (we can always enable INFO for StorageProxy
at runtime).  But you're right that people will want this info, which makes me think we should
not make them dig through logs but instead the return value should be map<version, list<endpoints>>,
that is, for each version list which endpoints had it.  (Missing endpoints could have '' version
or possibly null.)  So a map of size 1 would be agreement.

What do you think?

> Thrift interface call to query whether all live nodes agree on the schema id
> ----------------------------------------------------------------------------
>
>                 Key: CASSANDRA-1075
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-1075
>             Project: Cassandra
>          Issue Type: New Feature
>          Components: Core
>    Affects Versions: 0.7
>            Reporter: Benjamin Black
>            Assignee: Gary Dusbabek
>            Priority: Minor
>             Fix For: 0.7
>
>         Attachments: 0001-service-call-to-check-for-schema-agreement.patch, 0002-modify-system_-methods-to-return-UUID.patch
>
>
> Schema changes via the Thrift interface reach all nodes in a cluster at indeterminate
times.  It would be extremely useful to be able to query, via API, a single node in the cluster
and discover the current state of a transition to determine if all live nodes agree on the
schema.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


Mime
View raw message