cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jonathan Ellis (JIRA)" <>
Subject [jira] Commented: (CASSANDRA-213) complete bootstrap code
Date Mon, 03 Aug 2009 18:19:14 GMT


Jonathan Ellis commented on CASSANDRA-213:

this is the right thing to do though -- eventually, we want to have a --bootstrap-myself option
to start a new node, so it immediately goes into bootstrap mode (including waiting a minute
or so for a stable map) instead of participating immediately and serving up incorrect "no,
there's no data for that key" replies.

yes, this means it's slightly less straightforward to split a single existing node across
multiple new nodes in a single boostrap operation, but (a) the code isn't designed to do that
anyway so I don't see that as giving up much and (b) if we really need this it's easy to add
a delay on the existing node to wait a few seconds to see if more than one node wants to be
be a target.

the common op is "I'm adding a group of nodes mixed around the ring" so let's optimize for
that and make it as foolproof as possible.

> complete bootstrap code
> -----------------------
>                 Key: CASSANDRA-213
>                 URL:
>             Project: Cassandra
>          Issue Type: New Feature
>          Components: Core
>            Reporter: Jonathan Ellis
>            Assignee: Sandeep Tata
>            Priority: Critical
>             Fix For: 0.4
>         Attachments: 213-v1.patch, 213-v2.patch, 213-v3-addendum.patch, 213-v3.patch
> The bootstrap code sends files from DatabaseDescriptor.getBootstrapFileLocation to the
node that is joining the ring, but nothing actually puts files into that directory.
> There are probably other bugs / omissions we will find as we flesh this out.

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

View raw message