cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Sylvain Lebresne (JIRA)" <>
Subject [jira] [Commented] (CASSANDRA-4212) COMPACT STORAGE should not require a value to be aliased
Date Thu, 03 May 2012 15:46:50 GMT


Sylvain Lebresne commented on CASSANDRA-4212:

bq. apparently we write a copy of broadcast address as the value, but it is never read. Curious
now if that means there's a bug here.

No, adding the IP was some kind of anticipation that NodeId may be gossiped at some point
for the purpose of trying to shrink more aggressively counter context when they mention multiple
"old" nodeId. I figured if we did so we may end up storing id for other nodes and so the IP
was there to distinguish. But the idea I had at the time don't really work, so it ends up
being more or less useless information. In fact when I referred to having information in the
column I though (by looking to quickly) that this was the timestamp of when the nodeId was
created that we were storing (which is slightly more useful). But, that info is kept in the
timestamp in fact. Could be worth it to change that code to 1) start writing the IP as it's
unlikely we'll ever use it and 2) maybe it could be useful to add a simple way to query the
timestamp of columns in CQL, like say {{SELECT foo, timestamp(foo) ...}}.

Anyway, back on the issue, if we want to allow not defining a column alias for a system CF,
we can do. If we want to allow it to users, we'd need a little more code to allow insertion
in such CFs as said above.
> COMPACT STORAGE should not require a value to be aliased
> --------------------------------------------------------
>                 Key: CASSANDRA-4212
>                 URL:
>             Project: Cassandra
>          Issue Type: Bug
>          Components: Core
>    Affects Versions: 1.1.0
>            Reporter: Jonathan Ellis
>            Assignee: Jonathan Ellis
>              Labels: cql3
>             Fix For: 1.1.1
>         Attachments: 4212.txt
> It's legitimate to only need the column name in a schema, e.g., system.NodeIdInfo.

This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:!default.jspa
For more information on JIRA, see:


View raw message