cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Sylvain Lebresne (Updated) (JIRA)" <>
Subject [jira] [Updated] (CASSANDRA-3912) support incremental repair controlled by external agent
Date Tue, 03 Apr 2012 10:22:23 GMT


Sylvain Lebresne updated CASSANDRA-3912:

    Attachment: 3912_v2.txt

Attaching a new version (dubbed v2) of this. This changed the logic of getNeighbors a bit
so that:
# it returns no node if the provided range is not part of the local node ownership. Currently,
repair has been wrote with the assumption that the repaired range was a local one. It could
be that the code is able to cope with the range not being a local one (I haven't checked)
but I'd prefer avoiding the risk of foot-shooting currently.
# it only allows range that are contained in a local range. Which excludes ramge that overwrap
2 local ranges. I've explained the reason above: repair will repair data that doesn't need
to be otherwise.

Those are not really limiting anyway, since the correct way to use this is to call repair
on subsets of the primary range.

That patch also exposes forceTableRepairRange and getPrimaryRange (modified from the preceding
patch to return a list of String instead of a map entry just because that's what other JMX
function do) on JMX. The patch does not include the nodetool counterpart. The vague reason
is that since using this does require some understanding of what's going on and to write some
sufficiently smart external tooling to work correctly, it feels like exposing through JMX
should be enough and exposing through nodetool may add more confusion than benefits. I don't
stand very strongly on that though so I could be convinced otherwise if others feel differently.

I'll also note that I mentioned the flush problem for information/reminder purposes only.
I think fixing this one problem is beyond the scope of this issue (which really just expose
what already exists internally).

> support incremental repair controlled by external agent
> -------------------------------------------------------
>                 Key: CASSANDRA-3912
>                 URL:
>             Project: Cassandra
>          Issue Type: Improvement
>          Components: Core
>            Reporter: Peter Schuller
>            Assignee: Peter Schuller
>             Fix For: 1.2
>         Attachments: 3912_v2.txt, CASSANDRA-3912-trunk-v1.txt, CASSANDRA-3912-v2-001-add-nodetool-commands.txt,
> As a poor man's pre-cursor to CASSANDRA-2699, exposing the ability to repair small parts
of a range is extremely useful because it allows (with external scripting logic) to slowly
repair a node's content over time. Other than avoiding the bulkyness of complete repairs,
it means that you can safely do repairs even if you absolutely cannot afford e.g. disk spaces
spikes (see CASSANDRA-2699 for what the issues are).
> Attaching a patch that exposes a "repairincremental" command to nodetool, where you specify
a step and the number of total steps. Incrementally performing a repair in 100 steps, for
example, would be done by:
> {code}
> nodetool repairincremental 0 100
> nodetool repairincremental 1 100
> ...
> nodetool repairincremental 99 100
> {code}
> An external script can be used to keep track of what has been repaired and when. This
should allow (1) allow incremental repair to happen now/soon, and (2) allow experimentation
and evaluation for an implementation of CASSANDRA-2699 which I still think is a good idea.
This patch does nothing to help the average deployment, but at least makes incremental repair
possible given sufficient effort spent on external scripting.
> The big "no-no" about the patch is that it is entirely specific to RandomPartitioner
and BigIntegerToken. If someone can suggest a way to implement this command generically using
the Range/Token abstractions, I'd be happy to hear suggestions.
> An alternative would be to provide a nodetool command that allows you to simply specify
the specific token ranges on the command line. It makes using it a bit more difficult, but
would mean that it works for any partitioner and token type.
> Unless someone can suggest a better way to do this, I think I'll provide a patch that
does this. I'm still leaning towards supporting the simple "step N out of M" form though.

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